zingar

Could someone explain to me: are the top chess engines still good examples of the bitter lesson? My very poor understanding of the field is that it’s very algorithmic and not very much like what I think of as AI. But maybe “what I think of as AI” is not a criterion one can use to assess whether the bitter lesson applies.

It’s layoffs if the demand doesn’t grow to meet the output

I mean, that’s like half an hour of tokens, right? (Half joking laugh/cry etc)

I legit thought that this was what was being sold here

Noooo Strange New Worlds was the best thing that ever happened to trek

My main agentic coding side project is something that I can't justify paying the apple developer license for. If I was an Android person or I didn't have to pay the developer license, maybe I'd just go for it.

This kind of thing came up from time to time in the years before LLMs too. Agents would start with something based on English and optimize it until it became unintelligible to researchers. That was often something the researchers would shut down because they needed to be able to understand the comms.

Would love to know what he thinks nowadays

I think I do agree with the central point but a point of order is that if you tell AI to fix the bug and then tell it to fix the next error you might still be building an important skill: learning how to make the AI get good results.

My unsupervised agentic setup has built something that I wanted for decades but never built, so I guess that’s an increase in capabilities.

I’m curious, what was the application of AI that helped the family?

This is exactly my experience. A month of excitement building something that I wouldn’t have had time to do myself, then something broke in the setup and my motivation didn’t extend to fixing it.

I’m only back at it four months later and I don’t really know what happened before, or why it’s working now, I’m just happy that I can scratch that itch again.

That’s a measure of lines of code, I suspect the parent is talking about what results those LOC create.

AI built me a 1.5k+ LOC react component which is probably a 15x increase on the file size I would have created, with negative impact on the project for those extra LOC.

I think that people probably tried to tell me about the amazing stuff, but I couldn't take it in. This created a very pleasant surprise for me when I found out about the good feels after the dread of just expecting everything I know to be upended with no reward.

Well, a tiny fraction less than half the American people, at least the first time?

Do the agents come preinstalled in the images? Or do they somehow use whatever I’ve installed locally? The former makes sense to me but then I’m wondering whether the sandbox images stay up to date with new releases of each image.

Were you following any patterns/standards/advice on what you needed to protect against? Anything you can point the rest of us to?

The HuggingFace headline was criticised for being misleading about the level of agency demonstrated by the agent. Is this headline misleading? Is there anything here that I should know that would help me sleep easier? Is the worst thing about this what a human could do with Mythos on a big budget? (still pretty frightening, at least one of these techniques would work on me)

What’s your point in relation to this article?

Wow that is dark. Personal experience?

Do you think that the program design adds a lot on top of pointing out the current patterns?

I might be anchored by working with humans, but if I saw a plan for humans that included function signatures and what calls what I would say that is way too detailed. As a result I don't put it in my plans for AI either.

Enjoyed most of this but unconvinced by the program design part. If I see an agent writing function signatures or listing which functions to edit in a plan that tells me that I’ve given it too big a vertical slice. I always delete the code guesses. The thing that writes the code must always spend some time discovering where/what to write or it won’t have the right context. Or put another way: “decide first, act later” always feels worse than act-learn-act.

It would be interesting to know what the guardrails are. That would help me with understanding how I can use AI content. For instance I asked Claude to help me draw a diagram to represent a software engineering concept for a public presentation and then I had to stop and think: am I about to just reuse something from a Martin Fowler or Kent Beck book without attribution?

I feel like I'm seeing a maths+AI change from "let's test the limits of LLMs by seeing if they can do useful math" to "LLMs can do useful math, now let's solve lots of problems!", or put a different way the goal has shifted from "interesting exercise for AI" to "making a big difference in math". Am I correct?

Are there practical applications of any these problems being solved? No judgement implied, I'm well aware that "no" only means "not yet".

Ah of course, I do have a blind spot with local models.

Now that you mention it, Claude has had trouble at times with balancing brackets in the elisp I have it write.

Yet I still had the idea that LLMs should be better at lisp than other languages.

A fascinating contradiction, thanks for pointing this out.

I haven’t seen an LLM generate syntactically invalid code in any language in … a year? So I don’t know if “more likely to be syntactically valid” is a good enough selling point.

But maybe I’m interpreting you too narrowly?

The 1500 line react component that my unsupervised agentic coding side project produced says that it can get worse than we have now.

org-gcal … bit of a pain getting the Google API config right but so far so good.

Where does “big model highly quantized” start getting worse than “smaller model less quantized”? Is there a general formula or is it just trial and error?