As someone who spent countless nights tweaking Edge Detectors (looking at you, Canny), morphology operators, etc., building models to recognize 10 handwritten digits, let me tell you: the current set of LLMs (even the smaller ones) seem like magic. I had never imagined a computer would do such things in my lifetime.
Exactly people can say whatever they want, but current level of LLM is AGI level to me. It is already on par with senior programmer if the instruction/prompt is right.
Once we have 1000 tps, i am sure robots etc.. will also start working like magic.
I don’t know, I am writing a modest 30 page paper with Fable and even after rounds and rounds of feedback and improvements there are so many things that are just plain wrong or weirdly out of place or just stupidly written that Fable 5.1 doesn’t seem to have any awareness of by itself that I don’t think it’s AGI, I think a human researcher can easily outclass it in writing and problem understanding. It definitely has super human capabilities but it lacks awareness or self reflection in my opinion.
For example it should be easy to tell it to not write a paper in the style of a clickbait SEO article or use all of its stupid hallmark AI writing patterns “it’s A, not B!” And a smart human that would be told that would be easily able to comply with that but the model needs to be told in a very detailed way and it seems to lack even basic capabilities to reflect on this, when explicitly given a sentence it will be able to rewrite it but otherwise it’s mostly blind to it. That’s to me a hallmark of it being overtrained on the specific tasks or problems so it appears very smart but once you go off script it still shows that it’s not a “real” mind.
Of course it’s amazing and has super human capabilities in many areas but if you honestly think it’s better than Einstein like some people suggest why can’t it write a simple “good” academic paper even after giving it specific examples and instructions.
Maybe that’s what makes these things dangerous, they have super human capabilities in some areas but apparently lack self awareness, taste and meta reflection abilities. The only reason people aren’t afraid more is that they don’t act in the physical world yet, imagine giving it a body, superhuman strength and letting it care for your child when it has a strong “urge” to comply with your exact request and little to no self awareness and human basic instincts.
Though what has a program that is really really good at edge detection have to do with AGI? The community just spent decades to perfect edge detection. That's great! But let your edge detector try to fry an egg and then tell me again that it's AGI.
I'm still here, nearly 50 years and counting. If you had asked me what I imagined AGI would look like back in the 90's, I would have told you "A system that can do everything we can: see, hear, think, do.". If you had shown me GPT-6 back then, I would have said "It looks like a really powerful program, but that's not really what I had in mind.". That's AI, but it's not quite general.
And then you'd ask it about an area you're knowledgeable in and realise it routinely makes stupid mistakes.
Or you'd ask it to add a new page to your website and shout at it to use your existing brand colours instead of inventing some and realise it's not AGI at all...
Origami design will be my personal test bed for the coming years.
It's objectively very difficult and technical, it's spatiovisual, it's artistic, learning resources for it are sparse and most just learn by the FAFO method, current AI sucks terribly at it, and it's not likely to ever be specifically targeted by benchmaxxers.
Scientifically useful physics simulations. Every model absolutely sucks at them.
Or, as someone else points out in another thread here, academic writing. It's one of the things newer models seem to have actually gotten worse at. Even when you give them detailed instructions on how to write and what to avoid, the "load-bearing", "A but not B" and journal-like writing make it in anyway, with the supposed AGI having no ability to reflect on how blatantly unacademic (and often unreadable) its writing is.
Likewise, it's very impressive and useful, but it is obviously not AGI to those of us from that era.
If anything, the fact that it is so powerful is almost a concern, because I think we are still way underestimating what these systems will be able to do when we give them more cognitive capabilities.
At the moment we are something like, having had great success with propellers and have promised we will fly to the stars.
People love to say 'this is the worse they will ever be', then extrapolate to conclusion that they will continue to accelerate at the same rate of progress of last few years .. it may, maybe, or we will hit a ceiling, might be a temporary one, could be 5 years or 50 years ..
Anyone that is not impressed by what ChatGPT or the likes are doing now is being either dishonest or is incapable of being impressed.
Only the translation and language understanding capabilities are enough to be impressed, and they are 2 year old already. Now, the AI do see, draw, speak, listen, think, work, etc.
Someone from the 90's would simply not believe that the AI would be a machine but would think for sure that a human is behind. The only odd thing would be that this human would both exhibit high intelligence and stupidity at the same time.
I believe so. AIs are shockingly good at a lot of domains, but there's still a lot of pretty basic stuff they don't really "understand" at a conceptual level and (currently) they can't learn to get better at them.
(obviously it might take years for me to get good enough at something, or if you set the "arbitrary" task as something ridiculous, but lets work in good faith here and think of something the average human could do after learning about it)
If we progress to the point where an LLM instance can meaningfully learn to get better at something overtime without retraining, then I will accept that is basically AGI. Right now, they still seem to be pretty boxed into their training, even if you can prompt them to act differently.
Compared to 2016, it can do a lot of things, but it still fails for example with recommending a setup for my Raspberry Pi to have a 4G connection with some parameters (I want to use as a gateway between VoLTE calls and SMS, and my SIP server somewhere else). It failed miserably. I bought stuff according to its recommendation which was more or less a waste of money, twice. With miniscule knowledge compared to theirs, or even hobbyists', I could figure out all the details at the end, and order something which really works, but only after I sit down for 4 hours, and dig through exactly what I needed, because LLM lied flat out what Sixfab 3G/4G HAT can do. (Of course, not just LLMs lie, SixFab lied about something else too)
Of course, it's a moving goal post, because we have no clue what general intelligence is exactly. But it's definitely not general yet. Now the goalpost is to achieve that kind of level of thinking which I did in that 4 hours. When it reaches it, we will find something else it clearly lacks. Until we can't. Then, and only then we reached AGI. Until you see comments, reviews, etc about things which it cannot do, until then it's not general.
True. "AGI" has also become a marketing term. Achieving AGI has become valuable, so companies will move the AGI goalposts, over and over again, so they can achieve AGI, over and over again.
If you came at it from the perspective of imitating what the human brain does, we now have a very very powerful speech center and short term memory, and vision catching up. The other parts are missing. I‘m sure that’s being heavily researched.
This is a perspective I hadn’t heard or considered before.
Often times common crime (petty/violent) is viewed as a totally different area as white-collar crime or corruption. But they’re related.
As someone who’s had bikes stolen and car broken into, I sympathize with the desire for surveillance (especially in San Francisco). But Flock has no safeguards against cops using it to stalk ex-wives or misusing a partial match.
Yea, that's the thing: I could almost get behind mass surveillance if it actually solved crime. Probably not, but almost. But it doesn't! Having a camera on every street corner doesn't do anything to solve crime if the police won't look at it and actually follow up. So we get the downsides of mass surveillance without any upside.
Same thing with things like "find my iPhone". Great in theory, but if your phone actually gets stolen and you tell the cops "Hey, my phone is at this exact address, I can prove it with Find My," they still don't give a shit and don't lift a finger to solve it, so the technology is in reality useless.
We need a police force who is actually motivated to solve crime rather than just find black people to beat and show up in force when someone protests something.
the cops were in the apartment where my phone was stolen when my phone was stolen and had a video recording of it getting stolen clearly showing the person who stole it who was known to the police because it was the person the police were called about.
I said: can you get my phone back? They said: use find my, then come to the police station. I did. The police said: sorry, this location only shows the building and not the apartment, so we can't enter the apartment and get your phone
> Same thing with things like "find my iPhone". Great in theory, but if your phone actually gets stolen and you tell the cops "Hey, my phone is at this exact address, I can prove it with Find My," they still don't give a shit and don't lift a finger to solve it, so the technology is in reality useless.
The phrase I constantly recite to people at this point is "Mass surveillance is not mass enforcement and the two are separate words and concepts for a reason." People don't seem to get it until there's an effort to point that out.
I'm not sure if you know what you're talking about? Flock cameras have definitely involved in detecting people who come into SF in stolen cars or with stolen plates, who proceed to go on sprees of breaking into cars and stealing things.
I really dislike Flock as a company, but their cameras absolutely do get used in the way they're publicly pitched.
Flock doing bad things does not justify coming into this thread and stating as fact things that are patently false.
Plates can be retroactively searched, backtracked, subjects followed back to other cars and tracked back to where they left from, or other cars that are repeatedly in the area prior to cars being stolen before being used for other crimes can be identified and tracked backward. A plate doesn’t need to be flagged prior to a crime. It’s the same way they can tell which tolls and which plate readers a vehicle has passed recently when running a plate, and is frequently used to identify or trip up drug runners.
Translation: you are wrong and seem to be making things up to fit your position, because this is exactly what people have a problem with — that this data is retained regardless of suspicion.
I don’t think that’s true; my understanding is that there are (currently opt in, I don’t it should be) monitoring that flags any such requests for administrative review and a number of officers have been fired because of it. It is no different than a cop running their own or someone else’s license outside of actual police business, which frequently results in administrative reprimands/write-ups because of the same types of monitoring.
You might disagree on the punishments or protections, and that’s not to say it couldn’t happen, but it seems false to say there are no safeguards.
I do not “support” Flock. I would have less issue if there was mandated data retention policies (by law) that limited to thirty days without a court order. Afterward, Flock should be welcome to perform mandated anonymization/obscure biometrics/obfuscate identifiers to retain anonymous data at that point. Either someone is suspected of being a criminal and a judge agreed it warranted access and retention, or they are an ordinary private citizen in public. The government’s position that because the government doesn’t own the company they can get data from at any time, it isn’t government surveillance, is…
Well, I think it’s just lying but that is an opinion.
I’ve been robbed, multiple times, I’ve had my bike stolen… once I was robbed in London, surveillance capital of the world. Do you think the cameras helped? Take a guess.
I don’t sympathise at all with anyone who wants the powerful to surveil the rest of us. It’s just another, powerful, tool used by the wicked people who seek power over the rest of us.
Not cheap for sure but it's all for fun! I have done some optimizations to try to get cost as low as possible; the final clustering actually uses Kimi K2 for this reason. More info on https://intheweights.com/about
I literally had to solve the "preview Office files in the browser" problem last week. I couldn't find a decent solution, so I ended up making a endpoint that ran the files through headless libreoffice on the server to convert them to PDF.
For PPTX and DOCX, this solution is slightly worse than libreoffice conversion (this does not appear to output highlightable text, while PDF conversion does).
However, the XLSX preview BLEW my mind considering this was AI coded. Really good, even interactive!
Great project. The last time someone did this idea well they got acquired by Microsoft. Clipchamp has since been enshittified, making them ripe for disruption. The wheel continues to turn…
Cool, I can take all my self hosted stuff and stick it behind centralised enterprise tech to solve a problem caused by enterprise tech. Why even bother?
They don't. I'm using Cloudflare and 90%+ of the traffic I'm getting are still broken scrapers, a lot of them coming through residential proxies. I don't know what they block, but they're not very good at that. Or, to be more fair: I think the scrapers have gotten really good at what they do because there's real money to be made.
For logging, statistics etc. we have the Cloudflare bot protection on the standard paid level, ignore all IPs not from Europe (rough geolocation), and still have over twice the amount of bots that we had ~2 years ago.
The scrapers should use some discretion. There are some rather obvious optimizations. Content that is not changing is less likely to change in the future.
I think the point of the post was how something useless (AI) and its poorly implemented scrapers is wrecking havoc in a way that’s turning the internet into a digital desert.
That Cloudflare is trying to monetise “protection from AI” is just another grift in the sense that they can’t help themselves as a corp.
Only for in app purchases of Robux. They are uniquely allowed to distribute their own app store (not based on HTML/JS applets) within the App Store, which is against the terms that every other Apple developer agrees to. And they are allowed to use a virtual currency that can be obtained elsewhere without an Apple Tax to pay for digital goods purchases inside an iOS app, bypassing the IAP system, again against the terms.
reply