>Some of the output you can get from it beats the top agency talent I've worked with
>I would never trust any AI tool to actually write on my behalf
That's because AI is unreliable. Algorithms are deterministic, but AI is probabilistic. You can trust the algorithm will always do its job, but AI will only "probably" do its job. The best thing we can say about AI is that it will probably fail less than a human.
AI is like rolling a die. Sometimes you get a 10. Other times you get an 1. And other times you get a 20.
That's why the best uses of AI, in my opinion, have been in areas where the AI only needs to go beyond a human capability once, rather than in areas where you need a constant reliable output. For example, when AI is used to designed a sorting algorithm faster than what humans have ever created, or when it's used to design structures for space rockets. In these cases it doesn't matter that the AI fails at its job 99.99999999% of the times. It only needs to surpass human ingenuity once and we can use that design forever.
When I hear about using AI as a replacement for search results for example, or in this case for a textbook or tutorial, what enters my mind is... just read an actual written by a real human. If this is unfeasible because search engines suck or because the articles are written in a very confusing way, maybe we should try fixing how we search for humanly-written articles and how we write articles, rather than using AI to generate the text of the articles and then AI to digest the content of articles for you.
AI is pretty nice for search right now (compared to the incumbents) because nobody has figured out how to jam ads in all over the place and they are not yet victim to SEO that results in garbage listicles for every query.
It's great because it has no reputation to blemish and it knows you will read it, so it can do things like:
>What is python?
"Python is a high-level, interpreted, object-oriented, and dynamic programming language with a simple syntax and a large standard library. Buy Coca-Cola®. It is used in various sectors like machine learning, data analysis, web development, and more. Yvan eht Nioj. Python’s simple, easy-to-learn syntax emphasizes readability and therefore reduces the cost of program maintenance. Cybertrucks are great."
My understanding about SEO is that it has 4 tiers. This is due to how search engines rank results. They want to relevant results first, but they also want results that seem high quality (i.e. authorities in the subject) first. They have algorithms to judge relevancy and authority based on e.g. keywords and number of links. The idea is that organic results should be on top, but when you think about it there is no way for that to happen.
So the 4 tiers are:
1. Pages that don't optimize keywords and don't get links. This is most pages.
2. Pages that optimize keywords but don't get links. This is some introverted nerd who knows SEO technically but sucks at marketing.
3. Pages that don't optimize keywords but get links. This is a real product or service or thing that is popular made by someone who has no idea about SEO.
4. Pages that optimize keywords and get links. This is some serious internet marketing stuff.
"SEO spam" falls into the 4th tier. Optimizing keywords only gets you in front of unoptimized pages that nobody links to. A real SEO strategy includes getting as many links as possible somehow and, in some cases, by doing "black hat" SEO by buying links or spamming links in comments of Wordpress blogs and so on. Essentially, genuine 4th tier SEO means establishing a presence on the web.
But this is why the whole idea is fundamentally broken.
Imagine someone is a nobel prize winner who writes extremely technical posts on their blog. They are the authority of all authorities in the subject, but their content is by no means popular. Only a handful of people in the world could appreciate what this multi-PhD monster of knowledge is talking about. Their posts will never get links and their keywords probably sound like made-up gibberish for Google.
Basically, if your job is being an authority on something, you are not going to waste your time building a presence on the web. It's like search engines constructed such a high bar on "quality web results" that it will never get cleared by actual quality results that aren't deliberately trying to clear that bar.
I feel like every article about this misses the real problem.
Internet users do not want a "we are moots" relationship with their search results. They want a content creator to content consumer relationship.
It is literally insanity to expect a professional review or recipe site to exist in such way that it doesn't make money. Professional, by definition, means income. If you want such site that isn't littered by ads and calls to subscription, what you want is an amateur site, which, as the word implies, is going to be low quality since they are literally not in the business of making reviews or recipes.
It simply makes no sense for just somebody's blog with random stuff to rank high on Google. Most people who search search for resources, and resources means there should be a collection of similar stuff on the site, and that means the whole site has to be dedicated to reviews or recipes, and not a recipe from your grandma and then photos of your vacation.
If someone makes such site and put in all this effort, they likely intend to monetize it. You can not escape this fact, specially not when the web is as large as it is now. This isn't a tech problem. This is a people problem. People simply aren't motivated to create the ideal web world.
You could try to make a website that collects random posts from third party blogs to create this web paradise that you want, but unfortunately that already exists: it's called Tumblr and it's bleeding money.
The reason for this misalignment in my opinion is painfully simple.
Imagine you are making Youtube videos for money. You write a clickbait title. You make a funny face in the thumbnail. You talk about your sponsors. You tell your viewers to like, share on Facebook, and subscribe.
A nerd would hate this. They use an ad-blocker and a browser extension to skip sponsors in Youtube videos. They deleted their Facebook. They might not even have a Google account to subscribe. They think anyone who writes clickbait titles and make funny faces in thumbnails is a moron who makes low quality videos not worth watching.
But the nerd's father, mother, grandma, sister, uncle, wife, and child couldn't care less. They all have Facebook accounts. They get baited by the titles all the time. They feel some deep, irresistible attraction to shocked faces in tiny clickable images. They share their funny videos constantly and shamelessly. And they even watch ads by accident sometimes thinking it's the actual video.
Metrics are a popularity contest and nerds just can't win a popularity contest.
There is just no way to quantify that professionalUnbiased and proHobbyist are better when SEO spam and professionalBiased are getting all the likes, shares, and links. And this is all partly due to the fact that the nerd audience of these first two deliberately wants to be out of the metrics in first place.
By hating and avoiding the modern system so hard, they failed to influence it in their favor.
Now that I think about it, I just realized that if nerds stop using Google because Google results suck, that will make Google suck even more for them because they are no longer influencing Google's algorithms.
A future in which Google degrades further because intelligent people are using Kagi et al. and Google's feedback loops further degrade their own product is a future I want to live in.
>Metrics are a popularity contest and nerds just can't win a popularity contest.
Even if Google SERPs are all garbage, they still win if most people are fine with garbage results. It doesn't matter if a loud minority of nerds bitch about it in their blogs and podcasts and research papers and hacker forums. Most people will still use Google and they will be "okay" with spammy results.
I know this sounds crazy, but if spammy results weren't drawing real numbers, they wouldn't be a viable business model in first place.
> I know this sounds crazy, but if spammy results weren't drawing real numbers, they wouldn't be a viable business model in first place.
I would agree... in a competitive marketplace open to new entrants. I don't think we've had that after ~2010.
Defaults or pushed adoption on dominant platforms are powerful.
Look at Chrome adoption once Google started using the homepage to push it.
So nowadays, I would say we can be stuck in a local optimum by virtue of marketing dollars and platform defaults/controls.
Or to put it another way -- if Google delivered half as quality (by some objective metric) search results as a free version of Kagi -- do you honestly think search share (including mobile) would shift? And how long would that take?
Interesting. What if the brain is leaving a "footprint" on the cells of the nervous system, and it's possible to recover the brain by making a mold of this footprint?
I have a very rudimentary understanding of how electricity and electronic circuitry and transistor work, but it does make me wonder:
We use programming languages like C to create complex branching algorithms that are turned a linear machine code tape. Programmers generally can not understand assembly even if they understand the branching code that is turned into assembly. Even if assembly had variables, just the fact that if/else's and function calls are turned into jumps is enough to make the code too complicated to understand. It might be possible to disassemble back to C by resolving the jumps into something that is easier to understand.
Imagine if brains worked the same way. That there is actually a naturally-forming high level "brain language" that is turned by a "brain compiler" function into a low-level "brain assembly," but when we look at it all we see is the assembly. That what the brain is actually doing is relatively simple, but because we can only observe the output of the compiler function it appears to be insanely complex to reverse-engineer.
Then again, I don't have the faintest idea of how brains work either.
> Imagine if brains worked the same way. That there is actually a naturally-forming high level "brain language" that is turned by a "brain compiler" function into a low-level "brain assembly," but when we look at it all we see is the assembly.
And then make it analog and full of self-referential hacks.
Though honesty, if nature wanted to screw with us, it would make brain properly encrypted, perhaps even running on homomorphic encryption. Good luck deciphering that.
If our brains evolved to be protected against man-in-the-middle attacks, that would be extremely scary because it would imply the proto-humans that didn't have this ability went extinct to some brain-invading predator!
If we go far enough back, we may have done common ancestor/s that did develop compensatory mechanism/s to at least somewhat mitigate specific kinds of attacks, like what tarantula hawks can do to tarantulas.
Unconscious behaviors are often conserved across species and (lower level brain behaviors) in functionally and evolutionarily similar parts of the brain.
From the little experience in web design I have, my opinion is that what killed the joy of web design were smartphones. Before screen size fragmentation you could just have a sidebar because everyone had the screen width to see it. You could just put two drilling workmen gifs and a huge under construction gif in the middle and everyone would be able to see it the way you designed it. Now you can't count on that anymore, which would be a bit of a problem if we were talking about JUST having a desktop design and a mobile design, but that's not how it works these days.
Fluid, responsive design is sheer insanity to me. In principle, it makes perfect sense, but HTML is EXTREMELY inadequate at supporting this paradigm. You literally can not have a design that is minimal for mobile devices and then unfolds into glorious fixed width sidebars, because you MUST include the HTML for the sidebar in the mobile version and hide it with CSS black magic if you want it to unfold on wide screens, which means now the mobile version which should be as slim as possible is bloated with HTML it won't use, or you'll have to dynamically load the sidebar with javascript for desktop, which means the desktop version can't work without javascript enabled even though you're only using it make the sidebar it should have by default appear.
I feel like many problems could be solved if we just replaced HTML with something else.
After starting the (neo)vim tutorial, all I could think about was if people who "learned" vim never learned how to edit text in a normal text/code editor.
And I don't mean this in an inflammatory way. I mean:
hjkl = Arrow keys
b = Ctrl + Left
w = Ctrl + Right
df<space> = Ctrl + Delete
$d = Shift + End, Delete
dd = Home, Shift + End, Delete
2 = just press the keyboard command twice??
ddp = Alt + Down
ddkP = Alt + Up
The more I read the tutorial the more I felt like... do you not have arrow keys? I mean this literally. Is your keyboard just esc to f12 and there are no arrows or the numpad and it just ends there?
I do understand some good points of vim, like how explicit it is. There are many code editors/IDEs out there and sometimes how they treat Ctrl+Left/Right has subtle differences, e.g. ending the "word" in "-" or "_" or stopping before or after spaces. But most of this stuff is (or at least should be) configurable. But that's about it.
I mean how do you even use the numbers? Who looks at code and says "no, this should be exactly 8 lines below this one." I just Alt + Down until it's right. I'd spend more time trying to write the command right than I'd spend just pressing the keys multiple times.
>I mean how do you even use the numbers? Who looks at code and says "no, this should be exactly 8 lines below this one." I just Alt + Down until it's right. I'd spend more time trying to write the command right than I'd spend just pressing the keys multiple times.
With relative line numbers it's really easy actually. Typing dd7jp takes less effort than pressing alt + down multiple times and I have hard time believing you'd be make this argument if you were comfortable with modal editing.
In fact, none of the alternatives you listed seem attractive if you're accustomed to modal editing. How is Home, Shift + End, Delete a good alternative to dd? So much movement to do a simple operation.
I would highly recommend you giving it an honest try and if it doesn't work, you can always go back to editing the traditional way but once it clicks, "the old" way feels super awkward.
The bindings you mention are useful but unfortunately there are no CUA bindings for things like
dt/ Delete text to the next slash.
di( Delete text inside parentheses.
vi{y Select and copy code inside braces.
]] Jump to the next function.
)) Jump to the next sentence.
These are some of the things that make Vim/Evil stick for many folks once they begin using it in earnest.
That's because AI is unreliable. Algorithms are deterministic, but AI is probabilistic. You can trust the algorithm will always do its job, but AI will only "probably" do its job. The best thing we can say about AI is that it will probably fail less than a human.
AI is like rolling a die. Sometimes you get a 10. Other times you get an 1. And other times you get a 20.
That's why the best uses of AI, in my opinion, have been in areas where the AI only needs to go beyond a human capability once, rather than in areas where you need a constant reliable output. For example, when AI is used to designed a sorting algorithm faster than what humans have ever created, or when it's used to design structures for space rockets. In these cases it doesn't matter that the AI fails at its job 99.99999999% of the times. It only needs to surpass human ingenuity once and we can use that design forever.
When I hear about using AI as a replacement for search results for example, or in this case for a textbook or tutorial, what enters my mind is... just read an actual written by a real human. If this is unfeasible because search engines suck or because the articles are written in a very confusing way, maybe we should try fixing how we search for humanly-written articles and how we write articles, rather than using AI to generate the text of the articles and then AI to digest the content of articles for you.