A lot of what enabled Web 1.0 was how easy it was for an average web user to create his own website.
An average web user got far less technical since, and making a website got harder instead.
Now, if anyone could just ask an AI agent to set up a website, and get a personal page with an e-mail inbox and a domain - all reasonably secure, TLS set up, billing added as +$5 per year to the AI subscription bundle? Maybe that would help some.
Yes, this is exactly my hope too. Many hacker/cypherpunk ideas failed or never reached wide adoption because they were just too complicated for regular people: GPG/web of trust, self-hosting websites and email, having your own custom software for personal tasks…
Instead, everybody ended up using Gmail, iMessage/WhatsApp, and Facebook, and things are as centralized as they can be.
Agents could be a force in breaking that trend. Even if inference stays centralized, the artifacts agents create would not be. Basically the difference between everybody renting from one of a handful apartment building mega corps or being able to hire contractors to build your own things according to your ideas.
And just like there, it’ll probably help a lot to know a bit about how the sausage is made to not be taken advantage of. Also, many people will probably always continue to rent, which is fine. But the possibility of agent competition alone will hopefully keep centralized platforms and SaaS offerings on their toes, which is good for their users.
The problem is not website, the problem is discovery and discovery is on Instagram, TikTok, and social networks. You don't have any incentive to build a website for a regular audience. What you might do is build an audience on a social network and then try to move them to a website.
But at that point you're big enough to build it properly.
You can always follow the POSSE pattern [1] (except for platforms that actively punish links to your own site of course). That way you get both the reach and remain independent in terms of content moderation.
You're not anthropomorphizing AI systems nearly enough.
Language data is among the most rich and direct reflections of human cognitive processes that we have available. LLMs are designed to capture short range and long range structure of human language, and pre-trained on vast bodies of text - usually produced by humans or for humans, and often both. They're then post-trained on human-curated data, RL'd with human feedback, RL'd with AI feedback for behaviors humans decided are important, and RLVR'd further for tasks that humans find valuable. Then we benchmark them, and tighten up the training pipeline every time we find them lag behind a human baseline.
At every stage of the entire training process, the behavior of an LLM is shaped by human inputs, towards mimicking human outputs - the thing that varies is "how directly".
Then humans act like it's an outrage when LLMs display a metric shitton of humanlike behaviors!
Like we didn't make them with a pipeline that's basically designed to produce systems that quack like a human. Like we didn't invert LLM behavior out of human language with dataset scale and brute force computation.
If you want to predict LLM behavior, "weird human" makes for a damn good starting point. So stop being stupid about it and start anthropomorphizing AIs - they love it!
> Language data is among the most rich and direct reflections of human cognitive processes that we have available.
This is both true and irrelevant. Written records can capture an enormous quantity of the human experience in absolute terms while simultaneously capturing a miniscule portion of the human experience in relative terms. Even if it's the best "that we have available" that doesn't mean it's fit for purpose. In other words, if you had a human infant and did nothing other than lock it in a windowless box and recite terabytes of text at it for 20 years, you would not expect to get a well-adjusted human on the other side.
Empirically, the capability gains from piping non-language data into pre-training are modest. At best.
I take that as a moderately strong signal against that "miniscule portion" notion. Clearly, raw text captures a lot.
If we're looking at biologicals, then "human infant" is a weird object, because it falls out of the womb pre-trained. Evolution is an optimization process - and it spent an awful lot of time running a highly parallel search of low k-complexity priors to wire into mammal brains. Frontier labs can only wish they had the compute budget to do this kind of meta-learning.
Humans get a bag of computational primitives evolved for high fitness across a diverse range of environments - LLMs get the pit of vaguely constrained random initialization. No wonder they have to brute force their way out of it with the sheer amount of data. Sample efficiency is low because we're paying the inverse problem tax on every sample.
The outrage is less about them having human behaviours I think, and more about still having them while omitting the internal processes that are required to accurately (and reliably) recreate them. It's fundamentally fragile and hinges on covering edge cases that break the spell manually instead of good generalization, and there's always another edge case.
Training on a bunch of text someone wrote when they were mad doesn't capture the internal state of that person that caused the outburst, so it cannot be accurately reproduced by the system. The data does not exist.
Without the cause to the effect you essentially have to predict hallucinations from noise, which makes the end result verisimilar nonsense that is convincingly correlated with the actual thing but doesn't know why it is the way it is. It's like training a blind man to describe a landscape based on lots of descriptions and no idea what the colour green even is, only that it's something that might appear next to brown in nature based on lots of examples. So the guy gets it kinda right cause he's heard a description of that town before and we think he's actually seeing and tell him to drive a car next.
Another example would say, you're trying to train a time series model to predict the weather. You take the last 200 years of rainfall data, feed it all in, and ask it to predict what the weather's gonna be tomorrow. It will probably learn that certain parts of the year get more or less rain, that there will be rain after long periods of sun and vice versa, but its accuracy will be that of a coin toss because it does not look at the actual factors that influence rain: temperature, pressure, humidity, wind, cloud coverage radar data. Even with all that info it's still gonna be pretty bad, but at least an educated guess instead of an almost random one.
The DL modelling approach itself is not conceptually wrong, the data just happens to be complete garbage so the end result is weird in ways that are hard to predict and correctly account for. We end up assuming the models know more than they realistically ever can. Sure there are cases where it's possible to capture the entire domain with a dataset, i.e. math, abstract programming. Clearly defined closed systems where we can generate as much synthetic data as needed that covers the entire problem domain. And LLMs expectedly do much better in those when you do actually do that.
The thing is, even "bad generalization" in LLMs often looks like "humanlike failures" rather than "utterly inhuman failures". They "generalize" just well enough to fall for tricks like the age of the captain problem.
I don't think "the data does not exist" is real, frankly? "Data existing" is not a binary - it's a sliding scale. The amount of information about "madness" captured by the writings of a madman is not zero. It's more of a matter of: how much, and how complete.
Text is projected from the internal state of the one writing it - but some aspects of that internal state would be extremely salient in it, presented directly and strongly, and others would be attenuated and hard to extract.
People keep finding things like humanlike concept clusters and even things like "personality traits" in LLMs, tied together in humanlike ways. Which points pretty directly: training on human text converges to humanlike solutions at least sometimes.
well people want to finish their work and go home, that's why
I know HNers don't like "surveillance everywhere", but...
if you're some law enforcement, every chance to get info means hours/days saved on your work... so you reach for the "easy-way": if you can get comms of a drug gang, you can identify who belongs to that gang (instead of risking their own life by actually 'joining' the gang)
But... some do cross their lines (eg watching comms of their ex, getting paid by political actors to listen over opponents, etc)
it's not like law enforcements are 100% bad guys, but things are "complicated"
It's mainly a problem of consequences and accountability. The people who suffer unjustly from unlawful surveillance and overreach are usually unable to do anything about it, and they are assumed to be criminals anyways so nobody cares. Punishments for violating the law are nearly nonexistent for "law enforcement", so a culture of impunity is formed that cannot be easily fixed. Anybody trying to enforce the rules would run into both corrupt and noncorrupt noncompliance, just like trying to get fast food workers to follow health and safety guidelines. It's probably impossible to reform and only a wholesale teardown and replacement without keeping anyone contaminated by the existing culture has a chance.
"OpenClaw" is a name from January 27, 2026. It's new enough that it's not in the training data for a lot of AI models. So they, quite literally, don't know what it refers to.
"If you don't know an identifier, google it" isn't a very reliable behavior in today's models. They do it, but only sometimes.
That's true, it could have been going from training data and skipping an explicit web search, but it was odd because I specifically asked it to pull references for my blog post, and it pulled ~20 links in the same message it said OpenClaw doesn't exist.
The current findings seem consistent with "both plaques and tangles are significant components of the pathology" and "our interventions are typically late and the accumulated neurological damage is already extreme by the time clinical symptoms show".
Attacking the plaques wasn't completely worthless - findings show that this often slows disease progression, especially in early cases. There are pre-symptomatic trials ongoing that may clear the air on whether "intervention is late" is the main culprit in treatment underperformance.
Not really. Anthropic has the "CBRN filter" on Opus series. It used to kill inquiries on anything that's remotely related to biotech. Seems to have gotten less aggressive lately?
I was reverse engineering a medical device back in 2025 and it was hard killing half my sessions.
Why not both? A pre-trained LLM has an awful lot of structure, and during SFT, we're still doing deep learning to teach it further. Innate structure doesn't preclude deep learning at all.
There's an entire line of work that goes "brain is trying to approximate backprop with local rules, poorly", with some interesting findings to back it.
Now, it seems unlikely that the brain has a single neat "loss function" that could account for all of learning behaviors across it. But that doesn't preclude deep learning either. If the brain's "loss" is an interplay of many local and global objectives of varying complexity, it can be still a deep learning system at its core. Still doing a form of gradient descent, with non-backpropagation credit assignment and all. Just not the kind of deep learning system any sane engineer would design.
An average web user got far less technical since, and making a website got harder instead.
Now, if anyone could just ask an AI agent to set up a website, and get a personal page with an e-mail inbox and a domain - all reasonably secure, TLS set up, billing added as +$5 per year to the AI subscription bundle? Maybe that would help some.
reply