This has been the general consensus for about three years now. "Drastic increases in capability have happened the last 3-6 months" have been a constant refrain.
Without any data from the study past September I think its not unreasonable, if you want to make an argument based on evidence.
For me personally, I agree with you, I'm really seeing it as well.
> "Drastic increases in capability have happened the last 3-6 months" have been a constant refrain.
well, yeah. because that's been the experience for many people.
3 years ago, trying to use ChatGPT 3.5 for coding tasks was more of a gimmick than anything else, and was basically useless for helping me with my job.
today, agentic Opus 4.6 provides more value to me than probably 2 more human engineers on my team would
I tried GPT3.5 for translating code from typescript to rust. It made many mistakes in rust. It couldn't fix borrow checker issues. The context was so small that I could only feed it small amounts of my program at a time. It also introduced new bugs into the algorithm.
Yesterday I had an idea for a simple macos app I wanted. I prompted claude code. It programmed the whole thing start to finish in 10 minutes, no problem. I asked it to optimize the program using a technique I came up with, and it did. I asked it to make a web version and it did. (Though for some reason, the web version needed several rounds of "it doesn't work, here's the console output").
I'm slowly coming to terms with the idea that my job is fundamentally changing. I can get way more done by prompting claude than I can by writing the code myself.
There's a consensus that SOMETHING changed with Opus 4.5. It might have been the "merge rates" metric, it might have not.
I'm certainly getting faster and cleaner-looking solutions for certain issues on Opus 4.6 than I was 5 months ago, but I'm not sure about the ability to solve (or even weigh in) the actual hard stuff, i.e. the stuff I'm paid for.
And I'm definitely not sure about the supposed big step between 4.5 and 4.6. I'm literally not seeing any.
I don't know exactly how I would feel if the software I created selected a school to bomb and then suggested bombing the rescue parties trying to find / save any unexploded children 40 minutes later (double tap strategy to kill rescue parties and/or medics).
That 'let claude wing it, then send for review' approach that your lazy coworker uses is now how the largest military in the world operates. No big drama.
No country will be truly coal-free until they are a net energy exporter and they do not import any goods that use coal-based energy in their supply chain. Europe has de-industrialized which means it has effectively exported its coal burden.
Being coal-free is possible. Being fossil-fuel free is harder. Most of Irish energy comes from Natural Gas and Oil - the former is what supplanted Coal, not Wind.
This is a strange claim. During its peak years - in the mid 1990s - Moneypoint (the only coal plant in the country) provided 25% or more of the electricity mix while wind generation consisted of a few tiny pilot plants - contributing a miniscule.
In 2026, coal now provides 0% of the mix while wind provides 30% or more. Peat burning has also been fully phased out while oil (Tarbert) is in the process of being shut down while Moneypoint has been converted to oil but only participates in the capacity market - i.e. as an emergency/backup source - and so barely registers in the mix.
And even if coal was supplanted one-for-one with NG, it would still be a net win - by halving the CO2 intensity of generation as well as being far more flexible, scalable and much cheaper to deploy.
I’m not an expert here, but my understanding is that coal-free steel production is not a solved problem yet. And no, importing Chinese steel and moving the problem elsewhere isn’t a reason to pat yourself on the back.
There is absolutely no good reason to burn coal for electricity or heat in this day and age. If we had sane global leadership, every coal power plant left would be treated as a WMD and be bombed harder than that Iranian fuel depot.
Well, the term "natural gas" is ~200 years old and was invented to distinguish it from manufactured coal gas. But other than that, you're right. It's getting tiring having to explain to people why biogas isn't "natural gas".
Calling something a religion is a lazy and disingenuous way to dismiss it. The effects of airborne and waterborne pollution on human health has been the focus of scientific investigation for centuries at least. Ecosystem destruction is not "a religious belief" - it's something which has been measured, carefully recorded and studied using scientific methods.
> at 1/1000th the cost of re-wiring our entire energy system around renewables
This is alarmist hyperbole. Renewables now account for about 35% of the global electricity generation mix. There are countless examples of countries that run grids where renewables account for more than 80% of the electricity mix. Even if you just consider wind and solar, many countries are approaching 50% penetration in the space of 20 or 25 years without "re-wiring their entire their entire energy systems".
Yes NG generation is the only fossil tech still standing in terms of being able to compete at any level with modern generation tech but it is being squeezed.
Batteries surpassed NG in terms of economics for peaking sometime around 2020/2021 - which is reflected by the share of new capacity investment since then.
> industrializing 3B+ people using windmills
A windmill is a device used to mill materials that happens to be powered by wind, like a watermill, or a pepper mill or a paper/wood/etc mill. I'm not sure what that has to do with electricity generation.
I agree. Whenever numbers show that China is the largest CO2 polluter currently, it needs to be mentioned that China manufactures much of the world's physical goods.
- Peak coal usage is likely to be very soon in China (this year even according to some); after that coal usage flatten and start declining; all the way to a planned net zero in the 2060s.
The newer plants are designed to be more efficient, more flexible, and less polluting than the older ones. They are better at starting/stopping quickly/cheaply. Older coal plants used big boilers that had to heat up to build up steam before being able to generate power. This makes stopping and starting a plant slow and expensive. Because they consume a lot of fuel just to get the plant to the stage where it can actually generate power. The more often plants have to be stopped and started, the more wasteful this is. With the newer plants this is less costly and faster.
This makes them more suitable to be used in a non base load operational model where they can be spun up/down on a need to have basis. This is essential in a power grid that is dominated by the hundreds of GW of solar, wind, and battery.
What a lot of people also miss is that we’re in the age demographic bomb, where the global population is both aging rapidly and declining at the same time I.e. japanification
This means that global consumption will decline too which coincides with both factories and power plants shutting down
It kind of makes sense when you think we’re going to be running out of working adults like Japan, Germany, South Korea, Italy, and China to name a few countries. Even developing countries have seen their birthrates plummet. It’s like Children of Men IRL
Their existing grid uses coal because they have coal, just like the US uses gas because it has gas. And obviously as old coal plants are retired they're going to build new ones. They don't use the new plants for additional capacity. As they add more solar and storage, which they're building a lot of, they're going to absolutely crush the coal burning too. It's literally a national security issue for them.
This even compares cars built in an unrealistic 100% coal grid, and fuelled on 100% coal grid. Driving 14k miles a year for 10 years a Tesla Model 3 built under those extreme conditions beats a gas Fiat 500
As other posters below you have pointed out, it's not as simple as you make it out. You can't just stop building power plants overnight. The population and demands of China are growing and those needs need to be met immediately. There is no simpler, more understood way of rolling out new energy than building coal & gas power plants.
But look at the data. They are building clean energy solutions at a faster rate than any other country on the planet - by a huge margin. Scaling clean energy solutions is what we need, and it has to be done alongside the gradual phase-out of coal and gas.
Lots of people who don't read the HN Guidelines, apparently:
In Comments
Be kind. Don't be snarky. Converse curiously; don't cross-examine. Edit out swipes.
Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.
When disagreeing, please reply to the argument instead of calling names. "That is idiotic; 1 + 1 is 2, not 3" can be shortened to "1 + 1 is 2, not 3."
Don't be curmudgeonly. Thoughtful criticism is fine, but please don't be rigidly or generically negative.
Don't post generated comments or AI-edited comments. HN is for conversation between humans.
Please don't fulminate. Please don't sneer, including at the rest of the community.
Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.
Eschew flamebait. Avoid generic tangents. Omit internet tropes.
Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something.
Please don't use Hacker News for political or ideological battle. It tramples curiosity.
Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that".
Please don't pick the most provocative thing in an article or post to complain about in the thread. Find something interesting to respond to instead.
Throwaway accounts are ok for sensitive information, but please don't create accounts routinely. HN is a community—users should have an identity that others can relate to.
Please don't use uppercase for emphasis. Instead, put asterisks around it and it will get italicized. More formatting info here.
Please don't post insinuations about astroturfing, shilling, brigading, foreign agents, and the like. It degrades discussion and is usually mistaken. If you're worried about abuse, email [email protected] and we'll look at the data.
Don't feed egregious comments by replying; flag them instead.
Please don't complain about tangential annoyances—e.g. article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
Please don't comment about the voting on comments. It never does any good, and it makes boring reading.
Please don't post comments saying that HN is turning into Reddit. It's a semi-noob illusion, as old as the hills.
As other comments already point out, chinese coal power plants do not always operate under full load. They also decomission older more polluting ones.
Setting that aside, China has also dramatically pushed the electrification of their transportation sector like no one else. Considering BEVs and other electric modes of transport require less primary energy than fossil fuel equivalents, this checks out.
Supposedly they have been replacing old dirty coal plants with new cleaner ones alongside massive developments in renewables and nuclear. Getting air pollution controlled as fast as possible requires doing everything at once.
Coal is a lot cheaper and easier than modern energy sources when your goal is modernizing rural areas. Meanwhile, urban centers are decommissioning old emissive power plants and shifting to renewables. It's a fine way to do green transition and rural development.
By what measure? Coal hasn't been competitive for decades and the only way it had remained competitive in terms of cost per MWh even back then was if you burned it in huge (> 1GW) plants. 1GW plants are the very opposite of what you want to electrify rural areas - construction is slow and expensive, operating large plants requires considerable and qualified head-count, logistics is an issue and they require high capacity and expensive transmission systems.
And if your minimum unit size is 1GW then you lose the flexibility to roll out the tech incrementally - the average modern coal plant requires 3 to 5 weeks per year for scheduled downtime for maintenance - so your first 1GW coal plant requires a bunch of other generation sources to cover demand during these periods.
Solar and batteries are the obvious solution for rural electrification: scaleable, cheaper/simpler to deploy - no large scale civil engineering involved, trivial to "operate", effective without the support of big transmission systems and it's possible to buy everything off-the-shelf.
Coal requires transport and extraction which are both pretty expensive processes.
In my home town of ~300 people, there was just a couple of houses which used coal for heating. That's because sourcing and transporting coal was quiet expensive.
Electric heating was much more common. Even the old expensive baseboard resistive heaters.
When we talk about extreme rural areas, what you end up finding is solar and batteries end up being the most preferred energy sources. This has been true for decades. That higher upfront cost is offset by not having to transport fuel.
It's why you'll find a lot of cabins in pretty remote locations are ultimately solar powered. This is long before the precipitous price drop of solar.
How is coal cheaper and easier than buying and deploying solar panels and batteries. Both of which require basically zero additional infrastructure to deploy.
Last I checked mining and transporting coal required quite a lot of heavy industry equipment to do even vaguely economically.
If coal was cheaper and easier than other sources of energy, then the US would be building more coal power plants. But even with the Trump administration placing its weighty thumb on the scale to try and “save coal”. Coal plants are still being shutdown due to simple economics.
If existing plants can compete with renewables, to hard to understand how adding the cost of building new plants is going to change that.
It should be noted that the research from this article acknowledges that official Chinese coal consumption figures are often unreliable or subject to massive retrospective revisions. To compensate, the author uses "apparent consumption" (production + imports - exports +/- inventory changes) and power generation data. This methodology assumes perfect reporting of coal production and inventory levels across thousands of mines and plants. Historically, "statistical discrepancies" in Chinese energy data have masked millions of tons of CO2. If local provinces underreport coal use to meet "Dual Control" energy targets, the "flat or falling" trend could be a reporting artifact rather than a physical reality.
The article also notes that solar and wind capacity grew significantly faster than actual generation, suggesting "unreported curtailment" (where clean energy is wasted because the grid can’t handle it). If curtailment is rising, it means the "clean energy boom" is hitting a hard infrastructure ceiling. The research assumes that if grid issues are resolved, emissions will fall further. However, if the grid cannot integrate this power fast enough, the 290GW of coal power currently under construction will be called upon to fill the gap, potentially leading to a sharp emissions rebound in 2026–2027.
Further, a major driver of the emissions drop is the 7% decline in cement and 3% decline in steel emissions, linked to the ongoing real estate slump. This is a cyclical economic event, not necessarily a green technological victory. If the Chinese government pivots to a new stimulus package (e.g. massive "New Infrastructure" or high-tech manufacturing zones) to save GDP growth, the demand for steel and cement could surge again. The research treats the real estate decline as a permanent plateau, but China’s history of state-led investment suggests that industrial emissions can be turned back on by policy shifts.
Further, the analysis focuses almost exclusively on CO2. China is the world's largest emitter of methane, primarily from coal mine leakages. Even if CO2 emissions from burning coal are flat, continued coal production to feed the coal-to-chemical industry (which grew 12% in this report) results in significant methane venting. If you account for the Global Warming Potential (GWP) of methane, the total greenhouse gas trend might look much less optimistic than the CO2-only trend.
Further, the report credits a 3% increase in hydropower for helping suppress coal. Hydropower in China is extremely volatile and dictated by weather patterns (e.g. the 2022–2023 droughts). A single dry year in the Sichuan or Yunnan provinces can force a massive, immediate pivot back to coal-fired power to prevent blackouts. The 21-month trend may be as much a result of favorable rainfall as it is of solar panels, making it fragile and reversible.
Further, the report highlights a 12% growth in emissions from the chemical industry, driven by coal-to-liquids and coal-to-gas projects. This suggests that China is not decarbonizing its economy so much as re-carbonizing its industrial feedstocks. As China seeks energy security to reduce its reliance on imported oil and gas, it is building a massive coal-based chemical infrastructure. These emissions are harder to abate than power sector emissions and could eventually offset the gains made by wind and solar.
Ultimately, China is still the largest polluter (in every sense of the word) by a large margin. It's nice to see them taking steps to curb this, but we should all remember that any environmental benefits are purely coincidental to their goals of energy independence. We should expect them to rely on a diverse mix of sources - including the 8+ "mega" coal power plants coming online every month.
It should also be mentioned that despite being the factory of the world, China's CO2 emissions per capita are nearly half of the United States and comparable to some European countries.
> It should also be mentioned that despite being the factory of the world, China's CO2 emissions per capita are nearly half of the United States and comparable to some European countries.
To be fair, there's a large (~300mn) agricultural population in China who don't use developed country levels of energy. Nonetheless, this is still good.
Rural areas do not use much energy but Chinese cities are also more energy efficient per capita because of density and use of public transportation, walking, or electric mini scooters.
Europe is less industrial than in the past, but by every measure I can find many countries (especially Germany, Poland, Slovakia, Italy) are significantly more industrialized than the US - around 1.5x to 3x as much industrial activity and employment per capita, depending on the measure. Even the very least industrialized of the major EU nations (e.g. Spain, Greece) only just drop down to match the US numbers per-capita.
The issue is very complex. First - broad generalization - Europe's surviving industry is mostly made of less critical industries. If you look at important things in the world, and the important things that make up or make those important things, a tiny fraction of that is European, and that fraction is shrinking rapidly. There are some things - there is some green manufacturing stuff going on, there is some high-precision stuff in IT/CH/DE, there is ASML and Airbus, Poland can actually make things, etc. - but where will that be in ten or twenty years? I'll tell you: the high-precision stuff is rapidly moving to Asia, the green manufacturing is not very cost effective and uses a lot of imported core technologies, the C919 is going to fly with Chinese engines soon... the list goes on. The EU badly wants to make solar panels, cutting edge chips, fighter jets, rockets et cetera - and it simply can't, not at the cutting edge. The US, on the other hand, can make all of those things. It is still behind China in manufacturing overall, but it can still make a lot of the cutting edge, and it is still innovating.
Second, a lot of the EU stuff is already dead and only continues to exist through inertia. The median German cars and machine tools are worse than the median Chinese and they cost far more.
Third, those numbers often reflect the nebulous concept of "value added." Let's take the case of a refrigerator. Chinese company manufactures every technical part of the refrigerator and ships it to their EU business partner for €100. EU partner assembles it, fills it with foam, and sells it for €600. Most of the "value added" was in the EU! Win for the EU! Go EU manufacturing! The concept of "value added" is the basis for the entire EU VAT system and much of its economic indicators and incentives, while in the US it is almost never mentioned. This is also the source of the most hilarious comparisons (Greek manufacturing superior to the US per capita? χαχαχα)
If you want to cut through the bullshit, you have to look at actual things made. Among the US/CN/EU, who leads: Solar panels (CN), cutting edge chips (US), chipmaking equipment (EU), jet engines (US), aircraft (US), space launch vehicles (US), fighter jets (US), batteries (CN), nuclear reactors (CN), submarines (US), advanced missiles (US), cars (CN), CNC machines (CN), machine tools (CN), precision bearings and linear motion systems (CN), cutting edge medical equipment (US), gas turbines (US/EU), high voltage grid equipment (CN), telecom equipment (CN), construction equipment (US), ships (CN), advanced optics (EU), electric motors (CN), steel (CN), aluminum (CN), oil (US), cutting edge pharma (US), industrial robots (CN), wind turbines (CN), trains (CN), agricultural machinery (US/EU), drones (CN), smartphones (CN.) From that list, China leads eighteen, the US leads eleven, the EU leads two, and the EU and US are tied for two. And China is closing in fast on chipmaking. When China takes that crown, what will the EU have left?
Airbus seems to be leagues ahead of Boeing, not just in the civilian market, but also in military aircraft. Just look at their competing modern military tankers: the Boieng KC-46 is a worse plane than the Airbus A330 MRTT, but had huge cost overruns and delays.
EU is also at the cutting edge in helicopters, in fact 3 of the 5 new classes of manned helicopters introduced by the US military in the 21st century are from the EU: the MH-139, UH-72 and TH-73.
Submarines:
The Swedish Gotland and Blekinge class, and the German type 212 are both ahead of anything the US has. Wrt. bigger submarines, I don't think there's enough public information to argue that the French Triomphant-class is worse than the US Ohio-class
Advanced missiles:
The IRIS family, MBDA MICA and MBDA Meteor are cutting edge, European air-to-air missiles. MBDA also has a set of modern long, medium and short range anti-shipping missiles: the Otomat, Exocet and Marte. And that's on top of the evolutions of the Saab RBS 15 and Kongsberg/Diehl/Nammo 3SM. And the Swedish Saab NLAW anti-tank missile has been very successful with Ukraine in the past four years.
Cutting edge medical equipment:
Medical equipment is a huge field and very diverse and specialized, so it's easy to miss the areas where the EU is cutting edge. Just some examples I know of:
Siemens and Phillips are still top dogs in MRI machines. Three of the top five hearing aid companies are Danish: Demant, GN Store Nord and WS Audiology[0]. German Karl Storz is _the_ world leader in urology equipment. Danish Ambu is _the_ leader in single use endoscopes. Finnish Planmeca is a leader in dental equipment and their subsidiary Planmed one of the top 3 mammography companies in the world. Danish 3Shape and German exocad are more-or-less the only choices in dental implant CAD/CAM. Just to give a few examples
High voltage grid equipment:
Europe has been constructing a lot wind farms, many of them off-shore, and a decent amount of high-voltage, international electricity connections in recent decades. Most of that has been with European-made equipment. Some of the companies manufacturing that in Europe: Danish NKT, German Siemens and Swiss/Swedish Hitachi Energy (formerly ABB Power Grids) are three I know on top of my head. And then there are companies like Alstom that makes all the infrastructure around electric rail.
Ships:
European navies use warships built in Europe and I've seen nothing to suggest they are worse ships than Chinese warships. So the technology and shipyards are there to produce cutting-edge merchant ships, it's just not cost-effective.
Electric motors:
I've seen nothing to suggest Chinese motors have surpassed anything Swiss/Swedish ABB or Simens motors can do. And there are a ton of smaller, specialized motor manufacturers, e.g. Danish Grundfos that makes specialized motors for pumps.
Steel, aluminum:
The EU is self-sufficient in steel. A quick list of major companies producing steel in Europe: Spanish Acerinox, Luxembourgish ArcelorMittal, Austrian Voestalpine, German ThyssenKrupp, Italian Riva Group, Finnish Outokumpo, German Salzgitter, Swedish SSAB and French Vallourec.
Wrt. aluminum, the EU isn't quite self-sufficient. But ~75% of the imports are from Norway, Turkey, Iceland and Switzerland. So it depends on your definition of Europe.
Oil:
Oil is a commodity. You don't really gain anything technologically from producing it yourself, on the contrary it's seen as almost a curse, re:Dutch disease and so.
Cutting edge pharma:
If there's any category of company that's permanent fixture of EU stock indexes, it's pharma. To give just one example, Biontech, developer one of the two main Covid vaccines, is German.
Wind turbines:
In wind turbines Danish Vestas is number one and Spanish/German Siemens Gamesa is number two. The Chinese are catching up fast, but they're still behind.
Trains:
Spanish Talgo, French Alstom and German Siemens are all world-class EU train companies. Stadler is world-class, but Swiss, so it could also count. Then there's Hitachi Rail Italy (formerly AnsaldoBreda). As a Dane, I'm unwilling to call anything related to AnsaldoBreda "world-class", but the driverless trains they have supplied to the Copenhagen Metro meet the mark.
So I'd argue that there's at least 10 more categories where the EU is at least tied.
0: the last two are Swiss Sonova and American Starkey.
I think that determining a leader is usually far more clear cut. The EU and its bureaucratic-media apparatus love to find abstractions and subjective ways to discuss and measure things, but in reality you can measure production in objective terms. The EU is about 3-5% of global shipbuilding tonnage vs. China's 50-70%. Similarly, China produces 54% of the world's steel and 59% of the world's aluminum; the EU produces 7 and 4% respectively. China has 72% of the wind capacity installed and 4 of the 5 top turbine manufacturers. They similarly dominate numbers for trains (track laid, total HSR track, rolling stock built) and electric motors (market share of drone motors, motor components, EV traction motors, for industrial electric motors the EU and US are at about a quarter of the world market vs. China at 35-40%.)
There are a few areas where things are more subjective. Medical equipment, pharma and high-voltage, I could see a case for current EU ties and I appreciate your perspective.
Aircraft is certainly for the US, though. The EU really only has Airbus and Leonardo. Nobody would disagree that the US wins by far for: fighters, stealth, UAVs, transport, business jets, experimental, VTOL, long range bombers, and yes, helicopters. You cannot compare the US (over 2,000 Black Hawks in the Army alone) to the EU (100 Black Hawks) - the EU has a lot of lightly armed utility helicopters, the US has a massive array of everything from stealth, attack, logistical etc. Yes, sure, Leonardo can design a cool light utility copter, but that is infinitely easier and far less significant than an attack or stealth helicopter. The EU's attempts at serious helicopters (NH90, Airbus Tiger) are mostly a joke. Unless you want to count the AW129? War breaks out, how many of those can Italy send? ...4?
The picture with submarines is similar. The US has an expeditionary nuclear submarine force that can operate around the globe. The EU can do minor regional things, poorly. The Gotlands are cool, but they are small and the Swedes made three (3) of them - thirty years ago.
>Oil is a commodity. You don't really gain anything technologically from producing it yourself
The benefits in peacetime may be limited to dollars, sure; in wartime it is life or death. I am no huge fan of oil but in 2026 it still makes up most of our energy and most of the things around us. I also don't see anyone seriously accusing the US economy of Dutch disease.
>As a Dane, I'm unwilling to call anything related to AnsaldoBreda "world-class"
Hah, +1 for the chuckle :) I am Dutch... ask us what we think of AnsaldoBreda!
Somehow that’s an often missed aspect of this. Yeah, ditching coal has a wide array of nice side effects. It has killed many, many more than the world’s nuclear accidents.
It's worse than that, it's every 3 to 7 hours of fossil fuel pollution roughly equaling the total death toll of all nuclear power accidents in history (around 4000 indirectly, most from cancer resulting from Chernobyl - but there's only around 100 total in a direct way).
I think that depends on where you draw the line around the term "coal plant." There have been plenty of coal ash disasters that result in years of exclusion (for purposes of habitation, drinking water, fishing, etc.)[1][2][3][4]
Have you ever seen the common medical advice that pregnant women should avoid eating more than a few servings of seafood every week, and avoid certain kinds entirely, because they’re all contaminated with mercury? A huge portion of that mercury comes from burning coal. How’s that for an exclusion zone?
Nature would enjoy that. The economy not so much, depending on location. Around San Onofre (decommissioned now), a 30 mile Chernobyl-size exclusion zone would cover big chunks of Orange County and San Diego County. The US government recommended a 50 mile exclusion zone around Fukushima. 50 miles would cover southern Los Angeles and millions of people.
So The "worst case scenario" for nuclear power is creating a new wildlife park free from human interference [and emptying out half of Los Angeles]
If you look at net damage to the planet, fossil fuel burning energy sources kill literally 8 million+ people a year. Coal plants are vastly more radioactive than nuclear plants, and the effects of burning coal will have a vastly outsized share of damage to the planet in the long than nuclear. Its effects are just less concentrated to a single area.
And not all nuclear plants are the same. I don’t think it’s reasonable at all to compare Chernobyl to modern reactor designs, just because they both use the word “nuclear”.
Apso not sure if you are including coal mining, and all of the deaths and negative health outcomes as a result of the industry
Most of the exclusion zone is political nonsense. And overall coal has made much more areas much worse to live in. I rather live in the exclusion zone then next many coal plants.
Also there is a single case that happened from a non-western design. When looking at western countries like France, it shows how incredibly safe the whole industry is end to end.
Chernobyl's political nonsense was mostly down to the USSR wanting to deny that anything had, or possibly could, go wrong; if anything, the exclusion zone is the opposite of the western nonsense about nuclear power.
It's our unique freedom-themed nonsense, not the Soviet dictatorial-nonsense, which means we have radiation standards strict enough that it's not possible to convert a coal plant into a nuclear plant without first performing a nuclear decontamination process due to all the radioisotopes in the coal.
> When looking at western countries like France, it shows how incredibly safe the whole industry is end to end.
Relative to coal, absolutely. But don't assume western countries are immune to propaganda on these things, nuclear reactors are there for the spicy atoms, not the price tag or public safety.
Nuclear plans have objectively made power generation save and clean. When they were built in the 1960-1990s the were objectively the best and cleanest energy that saved a gigantic amount of lives.
The exclusion zone is nonsense because many that live in that zone has lower cancer rates then those outside. The idea is based on a invalid assumption about radiation an a linear relationship between radiation and harm. An I do think the standards we apply are to extreme in many cases mostly dating back to this misunderstanding about radiation.
As for the locality to nuclear plants and cancer, this is as far as I know been shown in many countries and as far as I know at least can mostly be explained by nuclear plants usually being built in industrial areas that often used to have coal plants and other industry going on.
> nuclear reactors are there for the spicy atoms, not the price tag or public safety.
Not sure what 'spicy' means in this context. In terms of price tag they are objectively a fantastic deal if built in larger numbers. Even in places where they were not built in the numbers they did in France, they are good life time deal, and give relativity stable long term prices.
And they don't have to be 'there' for public safety, they just need a good record on public safety and they do.
In places like Austria and Germany we have many known cases where a nuclear plant was planned and was prevented by activists, only to be replaced by coal, in both cases impacting 10000s of lives being worse financially in the long term.
What I didn't mention, in terms of propaganda, the anti-nuclear people are way ahead of any pro-nuclear propaganda. Its not even remotely close. The anti nuclear-weapons movement an environmental movement from the 1970s spread myths that are still repeated an often with emotional attachment.
My parents who lived in central Europe during Chernobyl hate nuclear power, while believing lots of nonsense that was in the news back then.
I have heard that the anti-nuclear propaganda is funded by other nuclear states, because of the aforementioned value of reactors (and their scientists and engineers) to weapons programs.
I have no idea if that's true, but it sounds very plausible.
However, if it was true, it would make it very much easier for governments to justify big spends on pro-nuclear propaganda. I mean, the USA managed to make test-detonations into a tourism opportunity: https://en.wikipedia.org/wiki/History_of_Las_Vegas#Atomic_te...
> Nuclear plans have objectively made power generation save and clean. When they were built in the 1960-1990s the were objectively the best and cleanest energy that saved a gigantic amount of lives.
Yes indeed. But that wasn't why they were built*, and that safety comes with an enormous cost. Ironically not including the exclusion zone, even with that included the amortised cost is quite small, but rather all the things that Chernobyl didn't have but should've plus all the international inspections and regulations to make sure nobody's secretly got a weapons program, either deliberately or via organised crime.
> The exclusion zone is nonsense because many that live in that zone has lower cancer rates then those outside. The idea is based on a invalid assumption about radiation an a linear relationship between radiation and harm.
That's a tiny sample size, consisting of people who are in poverty (and already mostly elderly) and can therefore be expected to have unusual health outcomes. Which could be higher or lower cancer rates, depending on what else is going on. Like, cancer won't get you if you pickle your liver too hard first with moonshine.
Though yes, the linear relationship between radiation and harm is known to be an oversimplification.
> Not sure what 'spicy' means in this context.
Radioisotopes. They're really useful for a lot of stuff, but the options for making them are mostly "fission plant" or "particle accelerator". This includes but is not limited to weapons.
> In terms of price tag they are objectively a fantastic deal if built in larger numbers. Even in places where they were not built in the numbers they did in France, they are good life time deal, and give relativity stable long term prices.
The cheapest are cheap, but the average and the trend line says they're mostly now a worse option than PV+batteries. Your milage, as the saying goes, may vary, so I wouldn't be even mildly surprised if e.g. Alaska says "hydro and nuclear" given this graph: https://en.wikipedia.org/wiki/File:Alaska_electricity_genera...
> In places like Austria and Germany we have many known cases where a nuclear plant was planned and was prevented by activists, only to be replaced by coal, in both cases impacting 10000s of lives being worse financially in the long term.
Indeed. It was terrible for the environment and general population health that we didn't have a huge roll-out of nuclear power up until the 2010s when renewables got interesting.
But unfortunately, although even the full cost (both monetary and to lives lost) of Chernobyl would, if amortised over all reactors, be much smaller than for fossil plants, the scale of that accident would have been an existential threat to many smaller nations. Arguably, it was existential damage to the USSR, even. People don't want to be subjected to someone else's game of Russian Roulette, not even when the overall odds of survival go up.
Because coal desposits in the ground have bits of Uranium and Thorium which are radioactive, they get concentrated in coal fly ash, and blow out the chimney in the smoke from a coal power plant, and kill people, they leach into the soil and waterways, and kill people.
That is, nuclear power plants only kill people by radioactivity in the case of an accident. Coal power plants do it in normal operation. As well as coal dust having a PM2.5 dust problem which kills people.
Make it about nuclear vs coal because people say coal is better than nuclear because it's not scary radiation, and it actually is.
> "Both are bad"
Nuclear generates more power from a Kg of fuel, with less CO2 pollution and fewer deaths. It's not bad, but even if it was bad it's not "both sides", it's much less bad.
Nuclear energy can be used to generate 24x7 energy as the grid-power to supply energy to a country whereas Solar and Wind require batteries.
I think that the last time I checked, when you take into factor the CO2 emissions and everything, Nuclear is the best source of Energy.
> I don’t think nuclear is the answer to things
I think that I am interested in seeing thorium based reactors or development with that too. That being said, Nuclear feels like the answer to me.
Feel free to correct me if you think I am wrong but I don't think that there is any better form of energy source than nuclear when you factor in everything.
Daytime doesn't mean the sun is out; the UK has heavy cloud cover and sunset near 4pm in mid-winter. https://grid.iamkate.com/ shows the UK is currently getting 10% of grid power from Solar at 3:30pm in March.
Sure. That's why there's the "interconnectors" section further down; the UK can take advantage of the fact that it's rarely simultaneously dark and zero wind across entire continents.
> Batteries are cheaper and faster to make in large quantities.
Yes I agree but their extraction at scale is still very C02 Expensive.
> No economy on the planet needs 24/7 peak power production. The times humans work correspond nicely with the times the sun is out.
With Nuclear energy, let's face it. If you have a nuclear plant running, the input is just some uranium which we have plenty of. Thereotically we have no problem with running at peak power production.
You are also forgetting that Sun can be blocked during times of rains and Wind is unpredictable as well.
If you can work with solar panels only that's really really great. Unfortunately that's not how the world works or how I see it function :(
You are forgetting that markets operate after work and the late night culture and so many other things. You need lights at energy and quite a decent bit. You are also forgetting that if we ever get Electric vehicles then we would need energy during late night as well.
A lot of energy in general is still needed during nights and would we be still burning coal for that?
With all of this, I am not sure why you'd not like Nuclear?
> You are also forgetting that Sun can be blocked during times of rains and Wind is unpredictable as well.
We already have wires that cross continents to smooth out supply variations. It's exceedingly rare you get no sun and no wind over entire continents for an extended period.
> You are forgetting that markets operate after work and the late night culture and so many other things.
Its not just about enough sun and wind capacity. There is already over supply in lot of the world. But the supply curve doesnt match the daily DYNAMIC demand curve so grid ops still dependent on coal and gas for different reasons. It becomes about what happens during unpredicted demand spikes, or when congestion on those wires happens whose load gets priority? which producers get curtailed? etc etc That moves probs into the political domain. You can watch daily grid ops live and see the probs. Wars and the weather randomly take down wires and substations all the time. If you can move people and factories to follow the wind and the sun then maybe you get demand and supply curves to match easier.
> We already have wires that cross continents to smooth out supply variations. It's exceedingly rare you get no sun and no wind over entire continents for an extended period.
I can be wrong but you would probably lose tons of efficiency even within High Voltage DC lines if everyday late night we take energy from different countries. Also this is getting outside of topic of discussion for me because one of the reasons we want Nuclear or Green energy in general is also the environmental plus the sovereign plus the long term affordability plans.
Another point from your first comment but if we run peak production in nuclear say in a country A, then the country A can also give power to Country B at late night similar to what you are proposed for solars.
> Again, batteries.
Once again, within my first comment I raise issue of battery. You mention a comment and I respond and then we get to batteries again.
I have no problem with solar at all without batteries but batteries really flip the equation in terms of environmental concerns.
My question is plain and simple, Why not Nuclear? I understand, I am not against Solar. Although environmentally, I feel like battery is a valid concern.
I am just saying that long term, Nuclear seems to be the better/best option. Why not Nuclear? That is a question which it seems that you may not have answered and that's a discussion worth having as well In my opinion too.
HVDC is more efficient than you think, 3.5% losses per 1000km. Which means intracontinental is obviously good, and intercontinental will work in some situations.
Nuclear power is expensive, enough that “what about night” is solve by building extra solar and batteries. Also, renewables wreck the economics of base load power that needs to run all the time to pay back loan, but can’t compete with solar during the day.
> You mention a comment and I respond and then we get to batteries again.
Yes. Because they're the answer here.
> Also this is getting outside of topic of discussion for me because one of the reasons we want Nuclear or Green energy in general is also the environmental plus the sovereign plus the long term affordability plans.
Good luck with nuclear sovereignty, if that's your concern. How many uranium mines are in the UK?
> Why not Nuclear?
/me gestures at the last 50 years of historical evidence
"Why not try nuclear" is like "why not try communism?" for physics nerds. We have tried it.
> Good luck with nuclear sovereignty, if that's your concern. How many uranium mines are in the UK?
I can't speak about UK but considering how cheap Uranium is, can UK not do cost analysis. Uranium is abundant material compared to Oil/Coal.
> /me gestures at the last 50 years of historical evidence
> "Why not try nuclear" is like "why not try communism?" for physics nerds. We have tried it.
Maybe, but I think that, I can speak about the problem within US which I can better explain but US had nuclear fearmongering attempts and Senators passed laws which increased regulations on it to the point that some regulations contradict past regulations.
Nuclear power plants being built on loan in such a flimsy regulatory market was what lead to the downfall essentially within US
Nuclear fearmongering and lobbying efforts from Oil Industry as they are one of the most strong opposers of nuclear energy[0]
Once again, how do I explain this but nuclear produces 3.2x less carbon emissions than Solar[1]
We are able to build hydropower plants, we are able to launch spaceships into moon and outer space. It's definitely possible to build nuclear if lobbying effort decreases.
I'd say that its our dependence on Oil and Coal which have been the problem. I have nothing against solar and that is something that I am saying from the start. At some point we should look towards transition towards nuclear as well. To give up on that would simply not be ideal.
> I can't speak about UK but considering how cheap Uranium is, can UK not do cost analysis. Uranium is abundant material compared to Oil/Coal.
Wait until you hear how cheap and abundant sunlight and wind are!
Economically useful uranium deposits are only proven in a handful of countries.
> We are able to build hydropower plants, we are able to launch spaceships into moon and outer space. It's definitely possible to build nuclear if lobbying effort decreases.
This is the "well we haven't tried real communism" argument again.
Alright, So I think that some/most of my talking points were very inspired from the michael moore's documentary on the topic and I re-watched it after reading your comments. (Although Michael doesn't talk about Nuclear in the shortcomings)
I was going to ask you 3-4 questions but then I searched them upon myself and I do think that the results are more (positive?) than I thought.
Solar could feed world's energy needs by 0.3% and I think that Excess Solar could be used for green Hydrogen etc. too when needed for burstable energy source and smart grids in general to fix the ramp-up/ramp-down problem
I think one of the only things that I was sort of worried about mainly was the fact that Batteries produce lots of Co2 emissions and harm to the planet when mined but it seems that they have lifespan of about 10 years and can be carbon negative 3-4 years.
I don't know, I go through waves of doubt over Solar. I might need to learn more about Solar because I feel like I can just agree to whatever side I hear the recent data's from. Lies, Damned Lies and Statistics it seems.
But I feel like although solar is right direction too, we probably need smarter grids and just improvements within grid infrastructure in general too. Another point about Solar could be that there can be a more personal adoption of it whereas I can't build my own nuclear power plant so I do agree with you.
I'd still say that there is a lot of greenwashing in the Climate Change community to treat wood-chips and trees as fuel source and all the problems that stem from that with timber industry.
So although there are short-comings in Solar given its intermittent nature. I do agree that unless Govt.s create nuclear, it could be a good bet for personal actions/ even Govt.s to diversify at the very least from Oil.
I still think that though there is something wrong where People are wrongfully worried about nuclear. India for example had 3% of its energy coming from Nuclear and I looked at wiki and we planned even more but anti nuclear protests started happening after Fukushima Disaster :(
I am still really interested about Thorium Reactors and the race towards building it though. They are mostly disaster-free and Indian in particular has quite a large reserve of Thorium (25% of the world supply). The govt. is working on making 100GW to raise thorium's ratio in energy to almost ~10% estimate from 3% till 2047 which would still be impressive given that total energy would skyrocket as well till then.
India has true chances of being Energy Independent long term if it focuses on nuclear and Solar both rather than focusing on Solar given any advancement in Thorium reactor will be huge for us. For reference Coal : Thorium power ratio for same mass is 1:3.5 Million and its even more efficient than Uranium.
Also Thorium cannot be used for Nuclear Bombs in the sense of a fission unless you drop it at someone complete point blank but at that point its worthless compared to missiles so we can genuinely share this technology all across the world.
Thorium Reactors long term feel the future to me. So maybe I am too bullish on Thorium.
Solar is nice but atleast personally, Investments in Thorium Reactors could make India Energy Independent given 25% of the supply. We also recently found a huge jackpot in lithium and other minerals in Kashmir recently so I suppose long term India can be sovereign in manufacturing batteries for Solar production as well.
There is such a massive possibility in nuclear especially more so for India and general consensus also being within Scientific community that nuclear energy is cleanest forms of energy. The Combination means that, I'd want my govt. to take some risks in nuclear research/projects given how big the reward can be and that's also why I vocally support Nuclear. Much more than Solar. But I'd say that any govt. has their own risk profile and maybe Solar can be boring but works technology for Energy Independence so I just hope that Solar & Thorium both show some good numbers long term as well. So it isn't as if I am anti Solar as much as I am very pro nuclear energy long term.
Respectfully, Can you tell me more about it because I genuinely don't know how you think Nuclear energy is bad. It's one of the cleanest forms of energy.
Is there any particular reason why you think Nuclear is bad in all honesty as its worth having a discussion here? Why do you feel Nuclear Energy is a hazard?
I understand if you feel Chernobyl or any event makes it sound dangerous but rather, Please take a look at this data on the number of death rates per unit of electricity production[0]
Oil is roughly 615x more deadly than nuclear. Nuclear, Solar and Wind (the renewables) are all less deadly and are 0.03,0.02 and 0.04 respectively and nuclear is a reliable source of energy source which can be used in actual generation.
Nuclear is very much a green energy. I'd like to hear your opinion about it.
Nuclear plants are great if they actually happen to get built and every person designing and operating them and storing the waste never makes a single mistake
> every person designing and operating them and storing the waste never makes a single mistake
Even within Chernobyl Disaster, it was a series of mistakes which led to the full scale disaster IIRC so it isn't as if a single mistake
Also Thorium based Nuclear Reactors wouldn't have this issue from what I understand as in the idea of explosions or anything,
> Nuclear plants are great if they actually happen to get built
I get this part but shouldn't this mean that people should be more vocal about support for Nuclear. We are vocal about support for Solar, might as well be vocal about support for Nuclear and Solar both too?
Not cost competitive with solar+batteries in many locales (less so the closer to the poles), and no learning curve, if anything a negative learning curve, nuclear never was more expensive than new nuclear.
And off course societal (and geopolitical) acceptance issues.
>And off course societal (and geopolitical) acceptance issues.
Right. One thing I've rarely heard emphasized is that, while nuclear power is not at all the same as nuclear weapons, it's still infrastructure that can be repurposed from one to the other. A world where nuclear is the predominant base load power source is a world where nuclear weapons are more accessible due to the proliferation of sibling technologies.
The cost competitiveness and societal issues make sense (though I suspect some of the cost is being externalized in terms of materials extraction and manufacturing).
I don’t understand what you mean by “no learning curve”. Do you mean that the learning curve is particularly steep for plant operators?
We make too few nuclear power plants for them to have a noticeable learning curve, and recently each subsequent one ends up more expensive than the latest, notably because of safety regulation. Korea and I think China had the best success in that regard (and France in the 80s) by being able to make real series, but you don't really see those now except maybe China.
It’s super expensive and it takes forever to build—so much so that fossil fuel companies fund “libertarian” voices to use it as an attack on environmentalists because nuclear means decades of unabated fossil fuel sales. If you commit to solar or wind, you start cutting into their business within as little as months.
The goal of net energy exporter assumes that energy produced at one time can be exchange for energy produced at an other time for the same price, and that assumption has not been true in Europe for decades. You can be a net energy exporter and still be dependent energy imports for more than 50% of the energy a country consumes, as has been demonstrated by Denmark.
I will happily trade 10 unit of energy for just a single unit of energy, assuming I get to decide when I give the 10 units and when I can demand the 1 unit. A lot of profit in the European energy market can be made by such a "bad" deal.
The date when a country energy grid is free from fossil fuels, like coal, is when the grid has no longer any demand during the year for producing or importing energy produced by fossil fuels.
Coal is essentially obsolete for energy generation. It's the most expensive and least flexible option. Even sticking with fossil fuels, natural gas is much better. It sticks around because many plants exist already, but new ones don't make sense.
It's even more nuanced than that because the United States is made up of many different states, with many different energy policies. Ireland would most closely equate to the state of Massachusetts by population and economic size, and Massachusetts shut down its last coal plant almost a decade ago.
I don't know if I buy this argument. If the US sends oil to another country, which burns it for energy, produces a finished good and exports that back to the US, then the CO2 released isn't accounted for in production-based numbers. But it seems to me like it isn't really properly accounted for in the consumption-based numbers that Noah is holding up, because those are effectively giving the US a credit for exporting the oil in the first place that offsets the imported good. As he says, the US's exports are carbon-intensive and that largely explains the difference being so small.
Noah also tries to refute the perception that manufacturing is in decline in the US, but he doesn't adjust per-capita and doesn't account for the obvious fact that major US exports are looking more and more like raw materials and less like finished goods, while imports are the other way around. Aircraft and ICs used to compete for top spot on the US export list. Since 2008 it's petroleum and oil.
But: EU is the only effective player in the world that drives energy policy outside its borders, by being a massive market with regulatory power regarding its imports.
If you look at three figures: energy use per capita, emissions per capita, and GDP per unit of energy/emissions, and include imported consumption, the EU's are all trending in a positive direction for many years now.
So stating the EU has de-industralized and its progress on shutting down coal is therefore 'fake' and misleading because it imports its industrial consumption from other countries to which it has simply offloaded its emissions, isn't true.
Great example of allowing perfect to be the enemy of good.
If major advanced economies are able to move their entire grid away from coal, it means the entire grid globally can move from coal.
"Ah", the critics say, "but manufacturing is so much more complex!"
Really? These are not countries without manufacturing. They have data centres stacked with the latest generation of Nvidia chips, electric rail, major capital cities, populations of millions...
... and of course, China agrees and is trying to move towards decarbonisation of their grid.
Yes, it'll take time, but it'll take even longer if you never start.
Coal is so deeply irrational. Only when you plug your ears and scream can you block out comprehension of the massive local externalities that make it inefficient compared to other energy options. It is cheap to setup with minimal access to highly skilled professionals so it was a good option to bootstrap economies until recently when solar, wind and NG have become easy to access and cost competitive. It's perfectly reasonable to have a phase out timeline to avoid under utilizing paid-for infrastructure, but it is a dead technology.
europes coal powerplants are in china, its polution is in china, the products of china are in europe and the producers from china live in europe and the us. China even offers greenwashing as a service, so people can buy for green notes a green consciousness.
> europes coal powerplants are in china, its polution is in china, the products of china are in europe and the producers from china live in europe and the us.
This is generally overstated. Emissions imported or exported via trade are significantly smaller than domestic emissions for almost every country. In the EU vs China case, accounting for imported/exported emissions basically changes which of the two is doing better, but emission levels are pretty close to begin with (US is already doing significantly worse than China either way).
For China, we are talking about ~1 ton/person/year from trade (in favor of China), while local emissions are at ~8 tons/person/year [1].
You make a valid point, but looking at the actual numbers it turns out that this makes (surprisingly) little difference.
This is what matters. The whole thing is an exercise in greenwashing. It doesn't matter if you stop burning coal in your own country, if the energy you import is also made by burning oil and gas.
The whole conversation about clean energy is polluted by the complete misunderstanding of the general population of how energy demands are balanced. Saying you're replacing coal and gas with wind is just nonsense. It's one solution to a bigger problem. The big problem is how to balance your grid across peaks and troughs and that requires a diverse set of clean energy solutions, with wind being one small part of it.
It's starting to become obvious that if you can't effectively use AI to build systems it is a skill issue. At the moment it is a mysterious, occasionally fickle, tool - but if you provide the correct feedback mechanisms and provide small tweaks and context at idiosyncrasies, it's possible to get agents to reliably build very complex systems.
> At the moment it is a mysterious, occasionally fickle, tool - but if you provide the correct feedback mechanisms and provide small tweaks and context at idiosyncrasies, it's possible to get agents to reliably build very complex.
This sounds like arguing you can use these models to beat a game of whack-a-mole if you just know all the unknown unknowns and prompt it correctly about them.
This is an assertion that is impossible to prove or disprove.
No it's more like if you knew how to build it before - LLM agents help you build it faster. There's really no useful analogy I can think of, but it fits my current role perfectly because my work is constantly interrupted by prod support, coordination, planning, context switching between issues etc.
I rarely have blocks of "flow time" to do focused work. With LLMs I can keep progressing in parallel and then when I get to the block of time where I can actually dive deep it's review and guidance again - focus on high impact stuff instead of the noise.
I don't think I'm any faster with this than my theoretical speed (LLMs spend a lot of time rebuilding context between steps, I have a feeling current level of agents is terrible at maintaining context for larger tasks, and also I'm guessing the model context length is white a lie - they might support working with 100k tokens but agents keep reloading stuff to context because old stuff is ignored).
In practice I can get more done because I can get into the flow and back onto the task a lot faster. Will see how this pans out long term, but in current role I don't think there are alternatives, my performance would be shit otherwise.
You could probably replace LLM with "junior engineer" here as it sounds like you're basically a manager now. The big negative that LLMs have in comparison with junior engineers is that they can't learn and internalise new information based on feedback.
I don't like that analogy. If I had to work with a Claude like junior I would ask for them to get removed from my team - inability to learn stuff, completely unexpected/unrelatable faliure modes and performance.
On the other hand Claudes tenacity, stamina and sustained speed is superhuman. The more capable models become the more valuable this is.
The assertion was "if you really know how to prompt, give feedback, do small corrections and fix LLM errors, then everything works fine".
It is impossible to prove or disprove because if everything DOES NOT work fine you can always say that the prompts were bad, the agent was not configured correctly, the model was old, etc. And if it DOES work, then all of the previous was done correctly, but without any decent definition of what correct means.
>And if it DOES work, then all of the previous was done correctly, but without any decent definition of what correct means.
If a program works, it means it's correct. If we know it's correct, it means we have a definition of what correct means otherwise how can we classify anything as "correct" or "incorrect". Then we can look at the prompts and see what was done in those prompts and those would be a "correct" way of prompting the LLM.
You don’t know it works. That you so glibly speak about products working is proof that your engineering judgment is impaired. You can’t infer the exact contents of a black box merely by looking at outside behavior.
The fundamental fallacy you are exhibiting here is similar to saying that rolling a six sided die and getting a “6” means that you will always get a 6 any time you roll it. And that if you get a 6 and wanted a 6, you must have therefore rolled those dice “correctly” and had you not gotten a 6 that would have meant you rolled them “wrong.”
>You don’t know it works. That you so glibly speak about products working is proof that your engineering judgment is impaired. You can’t infer the exact contents of a black box merely by looking at outside behavior.
I don't know the exact internals of a car. But I can infer my car works by driving it.
>The fundamental fallacy you are exhibiting here is similar to saying that rolling a six sided die and getting a “6” means that you will always get a 6 any time you roll it. And that if you get a 6 and wanted a 6, you must have therefore rolled those dice “correctly” and had you not gotten a 6 that would have meant you rolled them “wrong.”
Bro we rolled that dice MULTIPLE times. It's not a one time thing. And the "rolling" of the die is done with a CHAIN of MULTIPLE qureries strung together. This is not one roll. It's multitudes of data points. Yes results can be inconsistent from a technical standpoint, but the general result converges on a singular trend.
We know that much is true: a statistic and that is at most all we can say about reality as we know it as science formalized can only give a statistic as an answer.
"I don't know the exact internals of a car. But I can infer my car works by driving it."
No, you can't infer that it "works." Only that it CAN work. The car may be poisoning you with carbon monoxide. Your rear brakes may have become disconnected (happened to me). The antilock braking system may have a faulty sensor that only fails at very low speed, leading to them engaging when making a normal stop, but also preventing the mechanic from seeing the problem, because he didn't listen to your bug report and instead tried to repro the effect with high speed panic stops (also happened to me).
If I use a product and have a good experience, I can conclude that SOMETHING must be going well, but not that EVERYTHING is going well.
>No, you can't infer that it "works." Only that it CAN work. The car may be poisoning you with carbon monoxide. Your rear brakes may have become disconnected (happened to me). The antilock braking system may have a faulty sensor that only fails at very low speed, leading to them engaging when making a normal stop, but also preventing the mechanic from seeing the problem, because he didn't listen to your bug report and instead tried to repro the effect with high speed panic stops (also happened to me).
This is called pedantitic reasoning. You look like a drowning person trying to stay afloat.
Say i buy into your mysticism based take, is it a useful tool if it blows up in damn well near every professionals face?
lets say i accept you and you alone have the deep majiks required to use this tool correctly, when major platform devs could not so far, what makes this tool useful? Billions of dollars and environment ruining levels of worth it?
I'd say the only real use for these tools to date has been mass surveillance, and sometimes semi useful boilerplate.
> is it a useful tool if it blows up in damn well near every professionals face
It doesn't, that's ego-preserving cope. Saying that this stuff doesn't work for "damn well near every professional" because it doesn't work for you is like a thief saying "Everybody else steals, why are you picking on me"? It's not true, it's something you believe to protect your own self-image.
It works fine for me, because I don't trust to to do anything more than line changes, self-contained prototypes or minor updates. I've seen what happens when people give llms full control over authoring, and it has never worked out. Just look at the dumpster fire spotify has become. Its a buggy mess that barely works and is openly touted as being primarily llm developed at present.
point me towards something complex which llms have contributed towards significantly without massive oversight where they didnt fuck things up. I'll eat my words happily, with just a single example.
Honestly people are in such a weird place with this shit. I'm not saying don't read the fucking code - but I managed to get my setup to write 100k lines of indistinguishable SWE code in a week or so. The main limitation was my reading speed. This is something like a 10x speedup for me.
How does one verify 100k lines in a week? Let alone evaluate it to being SWE equivallent? That's super human. I like to think I am pretty good at what I do, but really critically engaging with 100k lines in a week is beyond even 10 of me. Forgive my skepticism, but I'm going to hazard the guess that you don't know what the fuck you're doing. You've lost your goddamn mind if you think you're doing anything other than skim read at a rate of 42 lines a minute for your entire work day without a break.
You have a fair perspective and I'm not going to try to move you from it. You have my exact opinion as of 3 months ago. I will just suggest you earnestly try it yourself.
I have, hence my perspective. For boiler plate and solved problems, it can do ok. However, anytime it needs to do something novel it will be hit and miss, usually miss. If it needs to do something where best practice has changed but training data overwhelmingly does bad / dated practice, you'll get that if you don't specify. Loads of issues like that. So, without giving small isolated and well framed tasks the output of which is then critically engaged with, you'll get increasingly opaque and unstable code.
On Saturday I had claude generate ~10k of lines of Lua code which uses the libASS subtitle format to build up nearly two dozen GUI widgets from subtitle drawing primitives, including nestable scrollable containers with clipping, drop down menus, animated tab bars, and everything else I could think of. I read probably about 100 lines of code myself that day, I "verified" the code only by testing out the demo claude was updating through the process.
Then on Sunday I woke up and had claude bang out a series of half a dozen projects each using this GUI library. First, a script that simply offers to loop a video when the end is reached. Updated several of my old scripts that just print text without any graphical formatting. Then more adventurous, a playlist visualizer with support for drag to reorder. Another that gives a nice little control overlay for TTS reading normal media subtitles. Another that let's people select clips from whatever they're watching, reorder them and write out an edit decision list, maybe I'll turn this one into a complete NLE today when I get home from work.
Reading every line of code? Why? The shit works, if I notice a bug I go back to claude and demand a "thoughtful and well reasoned" fix, without even caring what the fix will be so long as it works.
The concepts and building blocks used for all of this is shit I've learned myself the hard way, but to do it all myself would take weeks and I would certainly take many shortcuts, like certainly skipping animations and only implementing the bare minimum. The reason I could make that stuff work fast is because I already broadly knew the problem space, I've probably read the mpv manpage a thousand times before, so when the agent says its going to bind to shift+wheel for horizonal scrolling, I can tell it no, mpv has WHEEL_LEFT and RIGHT, use those. I can tell it to pump its brakes and stop planning to load a PNG overlay, because mpv will only load raw pixel data that way. I can tell it that dragging UI elements without simultaneously dragging the whole window certainly must be possible, because the first party OSC supports it so it should go read that mess of code and figure it out, which it dutifully does. If you know the problem space, you can get a whole lot done very fast, in a way that demonstrably works. Does it have bugs? I'd eat a hat if it doesn't. They'll get fixed if/when I find them. I'm not worried about it. Reading every line of code is for people writing airliner autopilots, not cheeky little desktop programs.
right, but we're not talking cheeky desktop personal programs here. Who cares what you do for your own devices and in your own time? Absolutely no one gets to tell you what to do there, so why discuss it? Live your life, do your thing, great! But that isn't what is being discussed here.
We're talking about the capabilities of the technology. I would not even think about using LLM output in life critical roles, but we're not talking about only such scenarios. I'm addressing your claim that these
> blows up in damn well near every professionals face
If you know how to use the tools, and know their limitations, you can generate vast quantities of useful code very quickly. If you can't manage this, its PEBKAC. You're saying you don't trust these to make more than minor changes, which might make sense if your code could kill people buy otherwise you're being overcautious or severely underestimating what these can do.
Useful is relative. Silly little toy aps? Sure go full auto, heck, i personally do. However, these will fail catastrophically or have failures that should never even have happened.
So, in professional environments, full auto is negligent to a point i hope it becomes a fireable offense. Like trusting lane-assist and adaptive-cruise in a car to handle full auto driving. It might even seem like it can, until the leading car disappears, until it hits a t intersection. You get me? Modern llms are lane assist and adaptive cruise, not full self driving. It frees up some of your headspace and attention but not all, in fact not even most of it.
> It's starting to become obvious that if you can't effectively use AI to build systems it is a skill issue.
I think it's fair to say that you can get a long way with Claude very quickly if you're an individual or part of a very small team working on a greenfield project. Certainly at project sizes up to around 100k lines of code, it's pretty great.
But I've been working startups off and on since 2024.
My last "big" job was with a company that had a codebase well into the millions of lines of code. And whilst I keep in contact with a bunch of the team there, and I know they do use Claude and other similar tools, I don't get the vibe it's having quite the same impact. And these are very talented engineers, so I don't think it's a skill either.
I think it's entirely possible that Claude is a great tool for bootstrapping and/or for solo devs or very small teams, but becomes considerably less effective when scaled across very large codebases, multiple teams, etc.
For me, on that last point, the jury is out. Hopefully the company I'm working with now grows to a point where that becomes a problem I need to worry about but, in the meantime, Claude is doing great for us.
I don't know if you deliberately cut-off the full point, but for the benefit of those with tired eyes I said 'feedback mechanisms', i.e. feedback in the control system sense.
We are still in early days and I'm sure this isn't the best way to approach this but here is what I do.
1. Agent context with platform/system idiosyncrasies, how to access tools, this is actually kept pretty minimal - and a line directing it to the plan document.
2. A plan document on how to make changes to the repo and work that needs to be done. This is a living document pruned by the orchestrating agent. Included in this document is a directive written by you to use, update the document after ever run. Here also is a guide on benchmarking, regression, unit tests that need to be performed every time.
2a. When an agent has a code change it is then analyzed by a council of subagents, each focused on a different area, some examples, security, maintainability, system architect, business domain expert. I encourage these to be adversarial "red team". We sit in the core loop until the code changes pass through the council.
2b. Additional subagents to create documentation, build architecture diagrams etc.
2c. A suggested workflow is created on how to independently invoke testing, and subagent, etc.
Would definitely tend to agree. Whenever I read complaints about accuracy of LLMs with complex systems, it has generally been from those that aren't thinking very critically about how they're using them in the first place. If you were to replace that LLM with a real human junior, would you really walk away for a few weeks and then assume the solution given was correct by default when you got it back? Obviously not. So you identify and gatekeep the most critical parts ahead of time, make error correction part of the process, and chunk the Giant, Complex Thing into Smaller, Achievable, Verifiable Things.
LLMs are proving to be very much force multipliers of the kind of developer you already are, and of those who report a 10x increase in productivity they're probably all being genuine. Whether that 10x is of careful, thoughtful choices or reckless rough-shod slop though is really an artifact of the developers themselves. I've been saying from the beginning that your effectiveness with LLMs is roughly equivalent to your ability to get effective results out of a real team of human contractors.
I'd agree, I've been building a personal assistant (https://github.com/skorokithakis/stavrobot) and I'm amazed that, for the first time ever, LLMs manage to build reliably, with much fewer bugs than I'd expect from a human, and without the repo devolving to unmaintainability after a few cycles.
It's really amazing, we've crossed a threshold, and I don't know what that means for our jobs.
No bugs means nothing if bugs get hidden and llms are great at hiding bugs and will absolutely fail to find some fairly critical ones. Your own repo, which is slop at best, fails to meet its core premise
> Another AI agent. This one is awesome, though, and very secure.
it isn't secure. It took me less than three minutes to find a vulnerability. Start engaging with your own code, it isn't as good as you think it is.
edit: i had kimi "red team" it out of curiosity, it found the main critical vulnerability i did and several others
Medium - 3 - DoS, Information Disclosure, Injection
You need to sit down and really think about what people who do know what they're doing are saying. You're going to get yourself into deep trouble with this. I'm not a security specialist, i take a recreational interest in security, and llm's are by no means expert. A human with skill and intent would, i would gamble, be able fuck your shit up in a major way.
it sure can help, but it shouldn't be considered solved. Realistically you should treat it as vulnerable until it's received some attention from natural human folks who do know what they're doing. LLM's are great but they're not experts in any field as of yet.
Right yeah, thanks for the constructive comment. Mind filing those vulnerabilities, or are you just making a point?
How do you know these are actual vulnerabilities? You just ran an LLM and it told you something and you came back to dunk on me, with zero context on the project.
Maybe you need to sit down and really think that you have no idea who you're talking to or what the project does. Next time you make a "omg this code is so shit" comment, include something more than "well my LLM says your LLM is bad" so we can have a discussion with facts rather than LLM-aided trashtalk.
EDIT: Out of curiosity, I've ran Kimi K2.5 on the codebase, and all the things it found are invalid, or explicit design decisions. So, next time you decide to tell someone their project "is slop" by running an LLM and relaying its verdict, consider a) the irony of what you're doing, and b) that the other person might know more than you about their own project that you spent "three minutes" running an LLM on.
The three minutes was my own time. There are thousands of projects like yours, why would i sink serious time into yours over another? I found one myself in three minutes, then ran kimi to see if it finds more. Someone being able to hijack your agent to run arbitrary code on your device seems like an odd design decision. The constructive feedbakc here is to take it seriously and engage with your codebase. Give a man a fish /Teach a man fish, you know?
Yeah except you didn't actually find a vulnerability, you just searched for a pattern, found it, and thought "vulnerability" because you aren't a security researcher and don't realize that context is important.
You should educate yourself more before you go around slandering people.
This can be done via a signal/telegram message, via compromised plug-ins, via file uploads etc. At no stage does there appear to be, as far as I can tell, any attempt to mitigate or limit attack surface. That's just one of the problems. If that's by design, my bad, you do you king.
Look, perhaps I've been to hard on you. Maybe you have an intellectual disability or something, I don't know you or your situation, my bad. I get that you're proud of your project, and I don't want to take that from you. By all means, explore, tinker, hack around and explore, these are great traits/qualities. Just don't tell people what you're offering is safe, or worse, safer than alternative offerings. That claim simply isn't true. To be direct with you, the 'how' of an attack isn't a riddle, it's a documented reality of the modern threat landscape.
You're hyper-focused on the front door, asking how an attacker would even message the bot, but you're ignoring the fact that modern attackers don't bother knocking. Read cloudflare's 2026 threat report, it's eye opening. Between automated session cloning and browser-based info-stealers (among many other modern headaches), the 'whitelisted user' is no longer a static, trusted entity. If a user on your list has their session token scraped via a malicious browser extension or a hijacked desktop app, the attacker effectively becomes that user. At that point, your bot doesn't see an intruder; it sees a 'trusted' account and hands them a loaded gun in the form of arbitrary SQL execution. Now, the problem is you aren't the only one with access to LLM's and obscurity never really was security, even less so now. An llm with credentials could easily probe its way through your bot's capabilities and connected data and exfiltrate everything.
So, the reason I'm calling into question your claims of having a 'safer' personal agent/bot/whatever is a matter of blast radius. A standard bot usually interacts with a restricted API or a set of hard-coded functions, so even if the account is compromised, the damage is capped. By giving an LLM the keys to the entire database, you've created a single point of failure that can result in total data exfiltration or a complete 'drop table' wipe, among any number of other nasty things. That's just _one_ issue in this project.
If you actually want this to live up to the 'safer than average' description, you have to move past the idea that a whitelist is a firewall. You need to distinguish between authentication and authorisation and implement defense-in-depth, starting with a database user that has zero permissions beyond simple 'Select' queries. You should be using a proxy that intercepts the LLM's generated SQL and kills any string containing 'Drop', 'Update', or 'Delete' before it ever touches your server, without some form of parsing/checking. Right now, you’ve built a powerful engine with no brakes, and telling people it’s safer just because it's on Signal is a dangerous misunderstanding of how modern exploits actually work.
Alternatively, fix how the project is described to be more accurate/honest than it is now.
which would be fine / legal, but brings us back to a question many turns ago, why invest time in this particular project? There's a sea of similar projects, what makes yours stand out as worth investing in? The claims made don't stack up, and there's no indication this is anything other than LLM output, which means there's nothing here that can't be re-created quickly to a similar level of quality, however you define that. If we start rewarding everyone who posts mediocre repos to HN with attention, the entire platform will get drowned in no time. Show some personal effort first, then I'll invest mine. Discourse on a public central page like HN at least benefits public discourse. Engaging with your repo? no real benefits as far as I can tell.
There is a grotesquely pulsing layer of overconfident dumbasses in business (and society in general) and this is the language they speak. My job at any company, as far as I can see it, is to make sure my local orbit is cleared of these wackos. They are parasitic extractors of value and soul.
Operational jobs are filled with these though. Not saying they are bad at work but the corporate culture leads to this language and style overwhelming everyone else and rising to the top
Children and animals can’t consent and are relatively helpless, so must be protected. Real incest leads to a high risk of developmental problems for subsequent children, so a ban makes sense.
Logically, none of these arguments apply to “step-incest”.
You say animals can't consent, but in the uk it's illegal for the ladies to have sex with male animals. That must imply some sort of consent. This also raises interesting moral questions for the farming industry.
(Your question doesn't make sense, but) A quick search shows that sex with an animal is definitely illegal in the UK [0] whether it's a male or a female human involved.
Physiologically, a male getting an erection implies stimulation, not consent.
However, more broadly I agree with you that laws protecting animals from harm can often lag behind what some might consider necessary. And despite laws protecting animals from "unnecessary suffering" [0] we still allow the slaughter of animals for food; presumably either because such slaughter is considered necessary, or because the suffering is considered reduced to an acceptably level.
I think you're missing the point. I stick my penis in an animal. That is illegal. I stick something else in there to get it pregnant. Not illegal.
So this isn't about consent. And it's not about the laws lagging. It's about the law differing depending on why you're doing the thing to the animal.
I don't doubt that some vegans might treat animal breeding as rape but I don't think the general population does, or consider that element suffering in any way.
Thanks for the interesting thoughts, your questions definitely makes sense to me.
"Children and animals can’t consent" view from _mst is a non perfect position but perhaps better that considering the opposite? Children can be groomed to consent and others animals training to. Here's a sensitive topic but I think it adds to the conversation: It has been reported dogs been used to rape humans. That's horrific for the victim, however consider this: sure the dog somewhat consent, but is it his own will if the trainer ask him to?
This is not to contradict you demonstration based on erection. Regulations for animals are sometimes justified, though they are extremely loose when tied to consumption.
I think the 'children can't consent' concept generally works.
I think the 'animals can't consent' thing is a retroactive justification for what is either ick, or a religious hangover.
I've had female dogs on heat try to hump me. Is that not consent?
We don't get dogs permission before breeding them.
The concept of bodily autonomy for animals is basically non existent.
The argument was originally that it was 'unnatural' but gay rights put the nail in the coffin there. 'consent' sounds good, but it isn't what's going on. It's the same 'unnatural' position, without the intellectual honesty.
Problem for me is. I kind of agree with the current cultural position. But there's no actual framework around it, and I don't like arbitrary rules for arbitrary reasons.
unnatural argument is to be ignored IMHO, it does't makes sense outside our emotional feeling and has multiple (and contradictory) interpretation so you'll end up arguing on each cases and don't have a rule/framework.
A thought experiment from Jonathan Haidt:
> Julie and Mark are brother and sister. They are traveling together in France
on summer vacation from college. One night they are stay- ing alone in a
cabin near the beach. They decide that it would be interesting and fun if
they tried making love. At the very least it would be a new experience for
each of them. Julie was already taking birth control pills, but Mark uses a
condom too, just to be safe. They both enjoy making love, but they decide
not to do it again. They keep that night as a special secret, which makes
them feel even closer to each other. What do you think about that? Was it
OK for them to make love?
I'm not saying unnatural is a good position. It is at least consistent with the attitudes we see.
Re the thought experiment. The standard argument is that of birth defects, which is obviously avoided. You could point to evolutionary pressures for the taboo also.
The thing is, kind of like pedophilia we've drawn a hard line. You could make the argument that someone the day before their 16th birthday is the same as the day after, therefore it's ok. I think the line being clear has more utility than the alternative. In this situation we're being invited to redraw a line, but where to put it? Do we make the getting pregnant illegal? Not being careful enough? So I would keep the line where it is.
Incest isn't just about genetics. People have adult godparents, foster children, adopted siblings and a whole load of other relations without genetic connections being involved. There are power dynamics and a whole load of other reasons it isn't right.
All very fair points; I was thinking (but didn't specify) of the step-brother/sister relationship as mentioned in other comments.
And just to be clear, I'm not an especial advocate here! I'm just interested as always in establishing clear logical rules that can then be applied fairly.)
I understand and agree with children, and bestiality being off-limits due to the non-consenting nature of any sexual interaction with them but a fantasy re-enacted by actors about incest/stepincest or even playing dead people? Why do you care? It's someone else's kink, not yours, you wouldn't like if the majority of people thought your vanilla sex kinks are off-limits (no more missionary for you, it's disgusting!).
People fantasise about a lot of things (not just sexual), but that doesn't mean they are right.
There are reasons beyond sex why incest (even with non-blood relatives) and bestiality are off the table. Not because of "vanilla" as you put it, but because they come with other costs. As does unprotected sex with multiple strangers, another stupid trope promoted in porn.
What reasons? You don't expand on it, you just say there are reasons why you agree on banning a type of fantasy to be produced between consenting adults, so other adults can consume and fulfill a kink.
So far nothing in that statement makes it comparable to bestiality or child porn, at all, why is that for you? Think answering this can move this conversation forward...
You seem to think that because someone has a kink or a fantasy that it automatically makes it okay, and we "shouldn't judge". If I consented before death for people to use my corpse for sexual purposes, than that does not make necrophilia healthy.
Don't forget driving, the most dangerous thing one can do in the West. It sounds be the first thing to be banned, really, the costs are just too enormous.
Western governments see car owners, smokers and drinkers as a steady income stream while preaching against them.
I was on a bus yesterday which took three times as long as it should have done. An hour and a half on a bus, and I still got in late despite setting out early. That was in a city. If you live outside a city, the public transport is even worse. I've missed job interviews, dates, other transport connections etc due to public transport. So yes, there are costs to that too.
"a single poorly preserved bone and radiocarbon dating"
You realize you are giving people the indication that you are knowledgeable on the subject, have read the article, or the underlying research - by framing your statement this way?
"much narrative archaeologists construct"
This comment has it's own narrative, to be skeptical of archaeologists. Be skeptical of the skeptics too.
reply