I've had the idea for years for a browser extension that showed positive/negative sentiment scores next to each post, based on comments. Mostly because often the top comment on posts are "ackchyually..."
I can't believe I'm saying this, but there may actually be a case for using "SI" now. I'm not saying machine learning somehow stopped being AI, but the term "AI" has become so synonymous with generative AI that people are apparently starting to retroactively redefine the entire field in terms of LLMs.
I prefer GenAI except the “Gen” stands for “general” because, really, the other kinds of ML/AI being contemplated in this thread are not at all general.
Like, before this you had extremely specialized models being painstakingly trained for an extremely specialized task using mountains of specialized data, painstakingly collected and labeled for correspondingly specific features, typically done by ML engineers (a specialized discipline) further specialized in that problem domain. And most users only have ever experienced all this abstractly through extremely narrow and specific functionalities like recommendations or ads.
On the other hand, any random person can hand these GenAI models any task that can be expressed as words or images or sound or some combination thereof and have a shockingly good chance of getting useful results. Yes they sometimes contain mistakes, but if you simply point them out, the models can often fix them!
I don’t know how much more general things can get than that.
It does get a bit uncomfortably close to “AGI”, which is a vague, heavily contentious term. But this sidesteps all that by not calling it AGI. The conventional “Generative AI” interpretation of GenAI thus becomes a convenient decoy for “General AI” ;-)
CI. Conversational interface. That's the difference. Simply a way to interact with a computer through the same mechanisms that you would use to interact with a person, through that computer's calculating a probable human reaction to a stimulus, after looking at millions of them and counting the likelihood of all possible reactions.
The real problem is pretending like there's any thought there at all. It's not programming because it knows how a computer works, it's programming because it can calculate what a person could likely write - which is why it writes source code rather than machine language. The value comes from the fact that it can chain a bunch of these reactions together, and react to things that don't originate from a human - if set to do so by a human, and they are things that a human could also predictably react to.
I'm not 100% sure you should be able to separate it from spam detection, though. It's learning from reading emails, and later whether that email is later marked as spam or saved from the spam pile. The critical part for me is whether that information is just going to be used to detect spam, or whether it's going to be used for something else.
If you're telling me you're not going to offer me spam detection, unless I not only agree to train your spam classifier, but some other model(s) too, just say that. It makes it clear that you're taking ownership of my data as payment.
As for children being able to sell their data as payment for the use of a search engine, the government pretends to be concerned about that - why doesn't it act here?
I am actually of the opinion that the problematic term is intelligence and that “super intelligence” won‘t actually help. This term has caused tons of issues in psychology and is now causing other issues in technology and journalism.
However I think it is a lost battle and I am willing to concede the term over to its current usage. I am however glad that traditional statistical models are no longer referred to by this trouble term. I think machine learning also has issues, and wish people would just be specific with their description by stating the particular algorithm they use (e.g. Markov Chains; ANN; Kalman Filter; etc.) but still accept “machine learning” this as a massive improvement over “artificial intelligence”.
Intel has supported native hardware decoder for AV1 since Tiger Lake. That’s quite old. You could buy a used NUC with Tiger Lake. Even when new a NUC with 1115G4 or 1135G7 isn’t expensive. I found one with damaged WiFi for $20. https://www.ebay.com/itm/366767313411 You can buy it now.
I went to the link to buy a computer (a NUC11TNKi3) with this capability for $20, but all it would let me do is place a bid on it; I cannot buy it now.
Cheaper to get a long fiber optic HDMI cable and connect your TV to your PC GPU. That's what I do. You get the best uncompressed video (typ 4K 120Hz or 8K 60Hz), zero lag so it can be used for stuff other than video, and don't have to maintain another box
I tried this and it was a huge pain. This works in a small place with a single TV and a single user, past that it's rough.
Running a hdmi cable through walls sucks, you're limited in tenancy (I can't game while wife watches the TV), controlling the pc remotely is annoying, most streaming apps don't support 4k on a pc (same problem with a nuc), I can go on.
This does not pass, or get anywhere close to passing the wife test imo.
Lots of media devices that don't have AV1 decoders tend to have issues decoding high resolution AV1 in real time so you either get pauses to buffer or dropped frames.
Nvidia Shield is hardly the most popular anything. Roku has 43% of the market and Amazon has 35%.
Would be interesting to see if we could segment third-party from TV-embedded (see quote), but the metrics from the article seem to reflect actual consumer usage, as most consumers just use what comes in the TV they buy nowadays.
> This reflects a consolidated approach and eliminates yet another device in the home, and it’s catching on, as just 34% consider streaming media players as their primary viewing device. Meanwhile, 56% said the same for smart TVs.
Very fair. Looks like Amazon is a bit better with AV1 support starting in 2021 with the Fire TV Stick 4k Max - 1st Gen[0], with a single absense of it on the seemingly uber-budget friendly $35 "Fire TV Stick HD Alexa Voice Remote (2024)".
That's still a sliver of the market, but the more royalty-free* codecs we use, the better.
>> The NVIDIA Shield TV (Pro and Tube) is widely considered the best Android TV streaming device for self-hosted media servers in 2026. While the Apple TV 4K wins on UI polish and the Onn 4K Pro wins on price, the Shield wins on something most other devices can't match: lossless audio bitstream passthrough. TrueHD, TrueHD Atmos, DTS-HD MA, DTS:X - the Shield passes them all to your AV receiver untouched.
>> Combined with hardware decode for H.264, H.265, VP9, and AV1, plus an 8-year-old Tegra X1+ chip that still gets driver updates, the Shield is a genuine "set it and forget it" Jellyfin client.
Nvidia shield is a device that simply works and passes the audio untouched to my sound receiver and I enjoy it __everyday__.
Nvidia hasn't released a new model of the Shield since 2019. The current Apple TV model, Google TV model, Roku model, Fire TV, and Onn models support AV1.
Shield is a 10 year old device at this point. It's essentially the 2017 hw that they've refreshed a couple of times with minor changes. Wait for the actual shield next for proper support.
The Shield is dead though and a long-time abandoned, right?
I remember the Shield going from fantastic to a horrible mess software wise over the span of a single Android generation (maybe 2019 or 2020), constant ARC weirdness, etc.
Sorry, but YouTube began thermal throttling on a Firestick the moment they started forcing AV1 regardless of the presence of on-device hardware support.
If the common hardware customers already have fails when you attempt to use a newer codec (but works fine otherwise), that can't be said to be a mainstream codec yet.
Your particular model is not representative of "the common hardware" as a whole. And it can be mainstream even when a significant fraction of devices can't use it. AV1 is a third of Netflix streams and most YouTube streams, that's more than enough to qualify as mainstream.
Even if half the devices were broken, it would still be more than a third of streams and mainstream.
A lot less than half of devices are broken when playing YouTube, I assure you.
If they stopped forcing it, it would still be the best choice most of the time. It doesn't "have to be" forced to be a major choice. The forcing changes the exact percentage but not the overall picture.
AV1 decode works well enough even on quite old devices, and it has more favorable quality and size, individually or when considering both. I prefer it to h264 (or vp9 or even h265) at this point, and there isn't a particular need for hardware decode to benefit from it. On Youtube (probably the biggest and most accessible source of av1 video), when comparing size and quality on different codecs (h264/vp9/av1), av1 is either gonna have much smaller size while keeping some level of quality, or have better quality at similar size, or even have both be better.
Firesticks are quite popular. I'd like to see what percentage in the wild actually does decode AV1 in hardware, and without stuttering or buffering problems. Even my newer devices sometimes choke on h265.
The first Apple chips with AV1 support only came out in 2023. So the majority of Apple hardware out there doesn't support it. It will be a long time till support is at >95%.
If you really want to see magic, take a test sample of YUV420 video and AV1 encode it to an extremely low bitrate at like 720p, 24 fps to simulate something like the jerky handheld front-phone-cam in the example URL. Keep lowering the bitrate until the quality become unacceptable. Then compare it to the same bitrate/approx total file size in H264 and how bad it looks.
At any point in time there's only a couple video codecs you should be using. When we can repeatedly shave off 30% of video bandwidth, there's no reason to stop making one or two new ones every decade.
Edit: And you can also look at the video codec landscape as having two major standards that put out new versions sometimes. The MPEG/h.26x series and the VPx/AVx series. It's not quite that simple but it's close.
There is no need to have a final one. It's no problem to keep supporting older codecs and add new ones in new products, as long as there's a "common ground" of widely supported codecs.
I don’t think we’re far off a “final” video codec, in the sense that jpg and mp3 and zip are effectively “final” image and sound and data codecs. Sure, there are newer codecs with higher compression foot a given quality level, but at some point it’s good enough to not bother upgrading, especially when the new version costs exponentially more for negligible gain.
…maybe THIS one will be The Final Codec… doodoodoo dooo
Re: mp3, Opus is a pretty big improvement. In my experience I've found that Opus is pretty reliably transparent at 96kbit/s, whereas LAME (which really squeezes mp3 for all it's got) needs ~130kbit/s
I think encode costs are higher (although I'm struggling to find any actual data) but in my experience this is totally negligible in practice.
Of course, don't go converting all your mp3s to Opus, but if you're running a music streaming server or whatever, Opus is pretty sweet... and probably close to "as good as it gets" until/unless we figure out some crazy new tricks.
> jpg and mp3 and zip are effectively “final” image and sound and data codecs
In popular consciousness, maybe. When people think image they thing jpg. But most websites you use aren't delivering images as jpg. Most sites care about saving money on bandwidth and their site loading fast, and because of that they tend to use webp and avif, which are much smaller for the same quality.
jpg: webp, avif, jxl. The first two are adopted because of non-negligible gains, the latter because of features missing in jpg.
mp3: Both AAC and opus have non-negligible gains and both are fairly widely used.
zip[1]: lz4, zstd, and lzma2 dominate at least gzip's implementation of deflate: For any given setting of gzip, there is a setting for at least one of those three that is both faster and smaller.
1: Not a codec; the zip format specifies a dozen methods for encoding. I'm assuming you mean Deflate, which is the most common compression method associated with zip.
Yes, "neural" or "learned" video codecs are considered the future. They already have competitive compression ratios, but last thing I heard was that they are still much slower than conventional codecs.
Trip planning is the most often used use-case for agents/bots I see in marketing. I'm lucky to travel once a year, but if I did have a trip, I wouldn't feel comfortable letting an LLM take the wheel. Too many variables for wrong flight or room, closed locations in itinerary, etc.
Now, critical server infrastructure? Sure, have at it Claude, make no mistakes!
I would literally never let an AI plan my trip. I do have very specific interests though and it's nice to have an overview of suggestions of places to go.
Plus, it's great at researching the history of places you're visiting, so you can find cultural experiences a bit easier.
I don't travel often, but I do live in Europe where it's easy to visit a new city or country in the weekend. This makes this functionality a quite common use for me.
I'm sure it will be a great resource for LLMs to read.
But seriously, good luck to them. I used to make Photoshop tutorials before YouTube, and it was a lot of fun. When YouTube came around, all the eyeballs went there, rightfully so -- videos made more sense for Photoshop. But I didn't have it in me to pivot.
And these days you don't even have to know how to do things like mask objects. There's one-click solutions, there's generative fill, light matching, face filters, etc. Hell, you can rotate a 2d layer in 3d now. Soon, live prompting in Photoshop will be the norm.
I used Kimi K3.1 Pro to figure it out as soon as cloudlandsdev put php-version.php live. The exploit worked just like this in a local Docker lab. But the terms page of the contest prohibited "resource exhaustion," so it stopped there. Showing Kimi the PoC, it says:
"The winning chain deliberately used per-request memory exhaustion, which appears to violate the literal no-resource-exhaustion rule, although the organizer accepted the win."
We're in October 2026. Updating packages is the easiest thing Claude can do. For me, I asked and received:
Node 26.3 → 26.12.1
Homebrew’s CA bundle, certifi and OpenSSL are updated, and so are Python 3.13 and 3.14 (3.13.17 and 3.14.9).
uv 0.10.9 → 0.13.5.
python.org Python 3.13: pip 26.2.2, installed from PyPI’s wheel with its checksum checked, then certifi 2026.10.19.
It also checked some of the other recent vulnerabilities like LiteLLM (already updated).
Sure, and by outsourcing this simple thing, you're just a little more complacent. Someday, when it does something silly, you won't notice.
It went OK this time. There's no guarantee that it'll go OK every time, for every thing.
All I'm saying is that understanding how to maintain the systems that you are responsible for maintaining is an important thing, and outsourcing that understanding is very dangerous.
"All I'm saying is that understanding how to maintain the systems"
Awwww. It is like talking to a dev in 2010. Ok. So, there are machines now smarter than you. You are the problem. You are the thing that adds errors. "understanding how to maintain" No, that is over. sorry. As a human, you are too slow, too dumb to compete.
When a CVE comes out, the patch needs to be applied across the company in minutes. YOU cannot do this 2010 dev dude. Put down the Puppet files. Let go of the python PR.
With this in mind, having LLM-powered NPCs would be a beautiful thing. There's been limited testing on PC games (PUBG Ally), but I can imagine it will be the standard.
> How many pelican riding bicycle SVGs were there before this test existed? What if the training data is being polluted with all these wonky results...
But Simon has demonstrated before that asking for any animal doing something silly will still get similar results from the same model.
Do you think this might be a kind of subconscious origin of your idea? A second of a move you saw in your early life / teens stayed with you as an example of an absurd thing to draw?
It's possible, but I think it's more likely that I was inspired by the TikZ unicorn thing, my love of pelicans (we have thousands of them in the harbor ten minutes from my house), and my knowledge that bicycles are difficult for humans to draw.
What about FOSS SaaS alternatives that are meant for team collab? Like Plane (Jira replacement). Does that have to be ran locally at home for it to be considered "self hosted"?
Self hosted means you host it yourself. I.e. you run the software. Running on your own home machine is definitely self-hosted. Hosting it on a VPS in some datacenter still is. Running it in a cloud platform somewhere is stretching it but I'd say that sill is. SaaS definitely is not self hosting even if the software is open source. The important part is that you're the one in control of the software.
reply