zlacker

I agree with your points but I'm also reminded of one my bigger learnings as a manager - the stuff I'm best at is the hardest, but most important, to delegate.

Sure it was easier to do it myself. But putting in the time to train, give context, develop guardrails, learn how to monitor etc ultimately taught me the skills needed to delegate effectively and multiply the teams output massively as we added people.

It's early days but I'm getting the same feeling with LLMs. It's as exhausting as training an overconfident but talented intern, but if you can work through it and somehow get it to produce something as good as you would do yourself, it's a massive multiplier.

>>avemur+(OP)
I don't totally understand the parallel you're drawing here. As a manager, I assume you're training more junior (in terms of their career or the company) engineers up so they can perform more autonomously in the future.

But you're not training LLMs as you use them really - do you mean that it's best to develop your own skill using LLMs in an area you already understand well?

I'm finding it a bit hard to square your comment about it being exhausting to catherd the LLM with it being a force multiplier.

>>avemur+(OP)
But... But... the multiplier isn't NEW!

You just explained how your work was affected by a big multiplier. At the end of training an intern you get a trained intern -- potentially a huge multiplier. ChatGPT is like an intern you can never train and will never get much better.

These are the same people who would no longer create or participate deeply in OSS (+100x multipler) bragging about the +2x multiplier they got in exchange.

>>conart+ra
The first person you pass your knowledge onto can pass it onto a second. ChatGPT will not only never build knowledge, it will never turn from the learner to the mentor passing hard-won knowledge on to another learner.

>>avemur+(OP)
Do LLMs learn? I had an impression you borrow a pretrained LLM that handles each query starting with the same initial state.

>>Goblin+jd
No, LLMs don't learn - each new conversation effectively clears the slate and resets them to their original state.

If you know what you're doing you can still "teach" them though, but it's on you to do that - you need to keep on iterating on things like the system prompt you are using and the context you feed in to the model.

>>johnma+la
Great point.

Humans really like to anthropomorphize things. Loud rumbles in the clouds? There must be a dude on top of a mountain somewhere who's in charge of it. Impressed by that tree? It must have a spirit that's like our spirits.

I think a lot of the reason LLMs are enjoying such a huge hype wave is that they invite that sort of anthropomorphization. It can be really hard to think about them in terms of what they actually are, because both our head-meat and our culture has so much support for casting things as other people.

>>Goblin+jd
Yes with few shots. you need to provide at least 2 examples of similar instructions and their corresponding solutions. But when you have to build few shots every time you prompt it feels like you're doing the work already.

Edit: grammar

>>johnma+la
No I'm talking about my own skills. How I onboard, structure 1on1s, run meetings, create and reuse certain processes, manage documentation (a form of org memory), check in on status, devise metrics and other indicators of system health. All of these compound and provide leverage even if the person leaves and a new one enters.the 30th person I onboarded and managed was orders of magnitude easier (for both of us) than the first.

With LLMs the better I get at the scaffolding and prompting, the less it feels like catherding (so far at least). Hence the comparison.

>>simonw+Ag
That's mostly, but not completely true. There are various strategies to get LLMs to remember previous conversations. ChatGPT, for example, remembers (for some loose definition of "remembers") all previous conversations you've had with it.

>>simonw+Ag
This sounds like trying to glue on supervised learning post-hoc.

Makes me wonder if there had been equal investment into specialized tools which used more fine-tuned statistical methods (like supervised learning), that we would have something much better then LLMs.

I keep thinking about spell checkers and auto-translators, which have been using machine learning for a while, with pretty impressive results (unless I’m mistaken I think most of those use supervised learning models). I have no doubt we will start seeing companies replacing these proven models with an LLM and a noticeable reduction in quality.

>>rerdav+8Z
I think if you use a very loose definition of learning: A stimuli which alters subsequent behavior you can claim this is learning. But if you tell a human to replace the word “is” with “are” in the next two sentences, this could hardly be considered learning, rather it is just following commands, even though it meets the previous loose definition. This is why in psychology we usually include some timescale for how long the altered behavior must last for it to be considered learning. A short-term altered behavior is usually called priming. But even then I wouldn’t even consider “following commands” to be neither priming nor learning, I would simply call it obeying.

If an LLM learned something when you gave it commands, it would probably be reflected in some adjusted weights in some of its operational matrix. This is true of human learning, we strengthen some neural connection, and when we receive a similar stimuli in a similar situation sometime in the future, the new stimuli will follow a slightly different path along its neural pathway and result in a altered behavior (or at least have a greater probability of an altered behavior). For an LLM to “learn” I would like to see something similar.

>>rerdav+8Z
I'd count ChatGPT memory as a feature of ChatGPT, not of the underlying LLM.

I wrote a bit about that here - I've turned it off: https://simonwillison.net/2025/May/21/chatgpt-new-memory/

>>runarb+8z1
I think you have an overly strict definition of what "learning" means. ChatGPT now has memory that lasts beyond the lifetime of it's context buffer, and now has at least medium term memory. (Actually I'm not entirely sure that they are not just using long persistent context buffers, but anyway).

Admittedly, you have to wrap LLMs to with stuff to get them to do that. If you want to rewrite the rules to excluded that then I will have to revise my statement that it is "mostly, but not completely true".

:-P

>>rerdav+Uw3
You also have to alter some neural pathways in your brain to follow commands. That doesn’t make it learning. Learned behavior is usually (but not always) reflected in long term changes to neural pathways outside of the language centers of the brain, and outside of the short-term memory. Ones you forget the command, and still apply the behavior, that is learning.

I think SSR schedulers are a good example of a Machine Learning algorithms that learns from it’s previous interactions. If you run the optimizer you will end up with a different weight matrix, and flashcards will be schedule differently. It has learned how well you retain these cards. But an LLM that is simply following orders has not learned anything, unless you feed the previous interaction back into the system to alter future outcomes, regardless of whether it “remembers” the original interactions. With the SSR, your review history is completely forgotten about. You could delete it, but the weight matrix keeps the optimized weights. If you delete your chat history with ChatGPT, it will not behave any differently based on the previous interaction.