jj4211

Well that description of a car is actually fairly close to the fundamentals, add an engine or motor and a steering wheel, and you've got it. Yes, a lot of engineering goes into the best possible realization of those basics, efficiency, suspension, safety, maintenance, and just a ton of more stuff, and it is a very valued execution above and beyond what, say, the Model T delivered. Automotive engineers have done hard and valuable work and complicated work, but no one is surprised that Model T led to faster, more comfortable, safer, more convenient vehicles that move around. It is a bit more surprising that LLM architecture works as well as it does while always focusing on the next token without ability to go further, best case running through and messing up and regenerating until you have operator appropriate output.

The 'seahorse emoji' was a pretty fun example of this at work, as it didn't have a seahorse emoji, but since it wasn't trying to generate the emoji, it started by building up the words to confirm and introduce the emoji since obviously there will be one, then putting up a wrong thing, then the words that would go after the wrong thing, but the weight still suggested there should be a correct answer and to start generating words for another try, and so on. "Reasoning" does the job of incurring this hit out of sight a lot of the time. Looking at the reasoning chains you'll see this behavior a fair amount, that the model suggests words that build toward an answer but fails on the key word and retries until something tests right or it models that it tried enough and it can't find the key word that would have been expected. It can of course digest it's own output and summarize the result without showing the operator spinning out, but it at all times is operating on the fundamental principle of model+very cleverly managed context influencing an answer one token at a time and ideally discarding the first run.

Monster3D... I was so freaking amazed when I got one... Nothing like the monitor flashing as that pass through switched the active graphics card....

lacks the ability to form a distinct new idea.

Yeah, but it's got that in common with a frighteningly large number of people...

See management, marketing, streaming, social media, etc.....

Actually, it's better at programming if you don't know how to program.

For example, someone who didn't know how to program sent me a patch to my code to 'fix' a problem they encountered. It was changed to silently swallow the error that really really would have needed to be fixed, but they were so enthusiastic that the problem looked to go away that they wouldn't let me know actually actionable debug information and just whined that I wouldn't take Claude's perfectly good fix for it. It is much better at satisfying folks that don't know how to look for better.

To satisfy my standards, I would have had to extract more debugging info, probably construct a test case tickling that exact situation making an expected error to ensure it won't just throw it away, and add it to the suite and then ask it to spin until the test case passed.

But alas, I'm neck deep in merge requests from folks that were formerly intimidated by coding and are now complaining that I'm not accepting more of their changes and more quickly because "all" I have to do is review the code which is obviously so much easier than writing it... It's like the "I have an app idea, all I need is for you to do it", but on steriods....

Particularly software development with very good and very quick tests allow rerolling and that is very appealing in those scenarios. Problems being that very good tests are rarer than people like to think and sometimes it just gets stuck in a loop. At work the other week someone set it at the task of fixing a bunch of build warnings that had accumulated over the years. It succeeded after burning through tokens to take 30 tries at it. It's solution after all that hard thinking? It put // @ts-nocheck at the top of every file and called it a day.

But superficially, someone handed it a chore and didn't have to think about it, and if no one looked deeper, then it was able to get to the desired behavior simply by rerunning the given task over and over without human intervention until it worked. Which is also broadly relatable as there's a lot of humans in the industry acting broadly the same, but I've always been frustrated by those folks anyway.

But that context is a mix of model output and other sources. The model output portion was generated token by token, and is combined in interesting ways with things like human response, search results, software output. It's still a backward looking mechanism, rather than having established a concept as a goal and then trying to build the words to reach that concept like we do.

Size and strategy for managing the context has been critical for improved subjective results, but it still doesn't exhibit the behavior of the words as a tool to address some concept, everything about the model is about the words themselves. So we end up with something very good at generating what seems right and there's a super high chance of what seems to be right actually being right. Especially when the software can automatically execute commands and the good or bad results reach into the context window, enabling it to effectively get automatically second guessed. The potential for automatic verification in some scenarios automatically feeding the context window is what makes it particularly appealing for software folks, though not universal.

I assumed that but everything I have seen as I dug deeper has been that at some level that is what is happening. If it is 'reasoning', it's generating a 'reasoning chain' next token by next token and using that to influence the final output tokens. The reasoning chain is discarded and since the actual output is a continuation of the reasoning chain it may conceptually be described as allowing the model to 'rethink' things, but even as the generation of a 'reasoning chain' has results that more closely resemble reasoning, it is still a scenario where it's building it one token at a time and we get to see meaning as an emergent property, rather than trying to find words to build to a more abstract concept like humans do. It just gets to throw away the intermediate work and the extra tokens manage to improve the 'accuracy' of the preserved final output.

I have a family member addicted to these (well, not Chinese...). It's pretty much all 100% slop, the voices, the animation, the script. My head hurts just seeing how it cuts every 5 seconds when I casually see it. The voices are so horribly soulless. The writing is inane and tortuously paced.

But people eat it up. Similar to (and large audience overlap) with people who toss massive money at shitty mobile games.

It’s worse than a human for most things since it doesn’t know truth from lie and will confidently say both as if they’re fact

It works for most executives and sales folks.

Baseless confidence is the recipe for business success, which is why they love these AI chatbots.

Bigger problem for the business leaders is how sycophantic they want to be to the user. If an insurance company used it for claims, it might actually approve a claim, and that would be unforgivable for them.

It's not even that it isn't 'ready for prime time', largely to the extent it works, it works with not so crazy requirements.

The problem is that they don't settle for what it is, they try to overextend it. In software development for example, in the cases where it works at all, you are 95% of the possible successes within 3 or 4 iterations. Problem is they demand that extra 5% which takes an order of magnitude more. Note that '100%' here is the max success possible with AI, not 100% success in general, that number varies greatly with context. So it might be in a certain scenario more like going from 19% AI curated to 20% AI curated, at huge incremental expense.

That's what is wild about it. At any given point in time, the model is wholly consumed only with the very next token. Maybe that token is a running narrative of 'reasoning' or directly in the output, either way, the AI does not have anything to model anything beyond the very next token. It doesn't have a destination in mind and is just finding the words to get there, it's building it up word by word. The overall 'meaning' is an emergent property of just picking the very next token and seeing what happens.

Honestly, it's shocking it works as well as it does. More shockingly, there are AI enthusiasts that argue that's how the human brain works, which I can't imagine someone going through life with every thought rooted in building it up word by word.

I had this happen to me once.

A woman asked me what was wrong with some technical thing and I explain that everything is just on crack and that's why it didn't work.

A bit offended, she declared she could handle a more technical explanation than that.

My response: "Well I can't"