elliotto

https://chatgpt.com/share/68e3674f-c220-800f-888c-81760e161d...

With thinking it spirals internally, runs a google search and then works it out.

GPT-5 Instant (no thinking) spirals wildly. Poor bot

I have the luxury of a delete button - the LLM doesn't get that privilege.

Much more important than citation farming a paper on 1 % improved performance

I strongly disagree with the authors political stance, which makes the fact that we agree on the problem / solution a nice bipartisan result

For example at the moment I have 1 long chat about some health stuff, 1 long chat about some bicycle repairs I'm doing, and then the rest are one offs. I rename the long ones (or just remember the name) and they stay at the top because I find them when I need.

I also manage memories assertively. If there's something important that I'm tired of repeating I just ask it to add it to memory and it will do it.

Most of my professional work is done in cursor so it's in a different place to my personal questions. These usually are one chat per feature kind of thing, and start a new one when it gets confused with the enormous context

This less wrong piece by a libertarian who examines the numbers and struggles to reconcile them with his beliefs is one of the best indictments on sports betting.

There is a small proportion of the population who cannot handle this. And they become prey to the predators in the sports betting industry. These guys make money off destroying their lives.

Embeddings had some context size limitations in our case - we were looking at large technical manuals. Gemini was the first to have a 1m context window, but for some reason its embedding window is tiny. I suspect the embeddings might start to break down when there's too much information.

You posit that the situation has a political cause, but I think this is just what happens when a system requiring exponential growth reaches the limits of its bounding box.

But it must be a wild ride to live while these cracks start to show. It probably looks like the greatest living mathematician making stream of thought tweets about how there aren't any small agents left.

I wonder whether this post is just a reflection of Terry living in California, which from where I sit looks like an end-stage capitalist hellscape.

Food distribution as not, and again in my country we are having constant investigations into monopolistic anti consumer behaviour by the large supermarkets.

I think about this as entropic decay - the lowest energy state of a social system (Hobbes' state of nature).

Lucky it is possible to create social and economic systems that use cooperation to produce better individual and group utility (dissipative structures). But there is a particular arm of politics that is currently dismantling these to feast on their cores.

https://inthesetimes.com/article/capitalism-job-bullshit-dav...

I'm not saying that these jobs are bullshit in the same way that a VP of box-ticking is, just that it's not a conspiracy that a cathedral based on 'design-doc culture' might produce incentives that result in people who focus on maximising their performance on these fiscally rewarding dot points, rather than actualising their innate belief in performant and maintainable systems.

I work at a start-up so if my code doesn't run we don't get paid. This motivates me to write it well.

Openai released study mode. I don't think it's anything special beyond a custom prompt that tells it to act as a teacher. But it's a good example of what these bots can do if you prompt them right.

The bots as they stand seem to be sycophantic and make a lot of assumptions (hallucinations) rather than asking for more clarification or instruction. This isn't really a core truth to bot behaviour, and is more based around adhering to American social norms for corporate communication - deference to authority etc. You can prompt the bots to behave more usefully for coding - one of my tricks is to tell the bot to ask me clarifying questions before writing any code. This prevents it from making assumptions around functionality that I haven't specified in the brief.

For non-coding use cases, I like exploring ideas with the bot, and then every now and then prompting it to steel-man the opposing view to make sure I'm not getting dragged down a rabbit hole. Then I can take the best of these ideas and form a conclusion - hegelian dialectics and all that.

I suspect history will remember this as a huge and dangerous mistake, and we will transition to an era of stoic question answering bots that push back harder

The current generation of LLMs have had their controversies, but these are still pre alpha products, and I suspect in the future we will look back on releasing them unleashed as a mistake. There's no reason the mistakes they make today can't be improved upon.

If your experiences with learning from a machine are similar to mine, then we can both see a whole new world coming that's going to take advantage of this interface.