xg15

But wow, they even want an account to read unlocked stories now...

If you just want to play with cool tech and make money, but for marketing reasons, you need to show some ostensible real-world reason why this tech should be developed, then "well, our chatbot could help lonely people with their relationship issues or give people in an abusive relationship a covert channel for help" is probably what comes to mind...

Runners up: Our AI startup will help the poor children in Africa and/or combat climate change (ignore the gas turbines, please).

It's starting with the exception before you even had a chance to understand the rule.

I guess this will become the frontline of the upcoming civil war in programming-land...

My point is that text-based interfaces - and voice is just a convenience around text - are not the best interfaces for a lot of tasks.

Unless you have a local model and the appropriate hardware for it, your agent is Somebody Else's Computer. Do you really want to send all your data and make your entire computing experience dependent on whatever OpenAI or Anthropic or whoever else is planning in this moment?

That's not even starting with inference time and token cost. Despite all the incredible advances in inference, it still takes more time than most non-AI computer functionality. Do you really want to wait a few minutes and pay money for something that you could also do with a few clicks fully locally on your PC?

But the most important thing: User interfaces. Right now, we're basically cramming everything you could possibly want to do at a computer into a chat interface. But there are lots of applications fir which specialized graphical interfaces are much more suitable. Why would you want to get rid of them?

Somewhat connected to that: Repetition. If you have to do the same task again every week or every day, it seems wasteful to ask the AI for it every time: Not just are you wasting a lot of time, energy and tokens, you're also at risk of getting inconsistent results, if the agent from today's session will interpret the requirements slightly differently than the agent from yesterday.

You can circumvent all those things by having the AI write you a custom app, but then using the app without AI to do the task.

Also, repeating a point from a similar thread: Software can have "intent" in the sense that it steers itself towards a predefined goal without having to be "alive" or "conscious" in any way. Some classic examples are thermostat control loops, navigation systems and chess engines.

So how to make the plugins remote-capable? You could write a massive virtualization layer that captures all system calls and forwards them to the remote - or, you could run the plugin on the remote and just pass the user commands and UI updates over the connection.

VSCode does the latter, so the nodejs runtime is where all the plugins are running on the remote.

(I understood the reason, I didn't say it was a good reason...)

Edit: another comment clarified it really goes both ways: The remote can use it to run code on the local machine as well, exactly what the "sandbox" pattern was supposed to prevent.

    Wander around the filesystem
    Edit arbitrary files
    Launch its own shell PTY processes
    Persist itself

Wait, could someone clarify which machine is being referred to here?

So in the author's setup, he runs VSCode (i.e. the front-end) on his dev laptop, which he wants to keep free of direct LLM access.

VSCode connects via ssh to a dedicated "sandbox" machine on which the LLM will be free to do whatever it wants (mostly).

VSCode realizes this the Microsoft way, by using the ssh connection to install VSCode Server on the sandbox machine - the "backend" - and communicating through it via a websocket connection.

So then, what happens? If the websocket connection allows the front-end to run arbitrary commands on the sandbox machine, this wouldn't be very exciting: The front-end already has an ssh connection and a massive server process that can do the same - and the entire purpose of the sandbox machine is to run arbitrary, untrusted commands without harm.

But the article says the websocket connection goes "back to your running VSCode front-end". So does that mean things are reversed? I.e. the agent/harness runs in the server on the sandox machine but for some reason has this websocket connection that also lets it run arbitrary commands on the dev laptop?

Is that it? That would be truly insane!

AFAIK, the first line is supposed to be the a very brief summary, while the other lines may contain additional information and can sometimes make up a pretty long text. Lots of tools make use of this convention and only show the first line if no detail information is needed.

Most tools that show the history only show the first line of each commit.

But I agree with you, this still assumes that long commit messages are rare and not that almost every commit has a huge message. Also, "long" doesn't mean you should put a novel in there.

Huh? I guess that depends on the exact definition of "classification", but I think the bulk of basic classification tasks we make every day to make sense of our surroundings, such as object recognition is definitely done using system 1. So is higher-level "stereotyping" or anything you could described with "I know it when I see it".

Because those responses can be incorrect or even harmful, you would sometimes make use of system 2 to correct them - but that doesn't change that the initial response is from system 1.

About Jev:

> We didn't train a model with Reinforcement Learning for Calibrated Decisions (RLCD) to calibrate the decisions and probabilities (even though they are not always correct).

Only 99% correctness! Borderline unusable!

About their model:

> It classifies: it gets a prompt with choices and outputs probabilities.

You want numbers, it gives you numbers! What more could you want?

If prompts are specifications for a change of the system's behavior, then it seems natural to manage them as changes and not as resources.

This would also keep them in the right "historical context" of the repo and avoid the "prompt rot" you were talking about.

But I don't really like the "LLMs as compiler" metaphor. If you followed that logic to the end, you'd have to "rebuild" your entire project from the spec every time the spec changes. Not just would the token cost be insane, but you'd also get a completely different implementation each time, maybe with different UI and design decisions where the spec left things open.

The alternative is to see the code as the source of truth and LLMs as (extremely sophisticated) editing or refactoring tools. Then by all means, still check in your prompts, but now they are documentation on how a feature was implemented, not the source of truth themselves.

(Or did it look up the results on the web?)

I think the order of concerns is also telling.

Primary reason why it's bad: Increase risk of runaway superintellicence, XRisk, etc etc

Secondary reason: Oh, it will also make most people unable to have a living wage. We should probably do something about that, if it doesn't strain the economy too much.

The dehumanization of everyone who will still have a job? Nah, not a problem.

Though I think this dismisses the potential of people to organize, especially in jobs like those. This is why I'm skeptical that future will happen as easily as he is warning here.

...and of course he is a Rationalist/EA guy. Why am I not surprised?

But I think even purely economically this won't work: For starters, who would set the price for your book?

Your book would become an exchangeable commodity, with the "purchase decision" done solely by AIs (and their companies), not the "readers". The financial success of your book would depend on whatever criteria AIs use to decide about including your content or not, not your (human) readers (or what's left of them). As with most commodities, this would enter a "race to the bottom" that would probably leave you with pennies - or steer your writing into a direction that you were never intending to go in.

Just a question: In such a system, why would you still write the book yourself and not have an LLM write it?

Yes, as kids probably always did since the invention of the children's room. The usual solution was to educate - and push - them until they hopefully started to understand the benefits that an orderly room can give you and maybe even had some fun thinking of a good ordering scheme by themselves.

Only in tech would the response to kids being too lazy to clean up their rooms be "oh well obviously keeping your stuff in order is an unworkable strategy".

I remember that, back when we were in school, we would actually spend a lot of time structuring the bookmark folder, making elaborate menu hierarchies. Doing this the first time was fun.

The problem was, this lasted about until the first time the PC had to be reinstalled or someone switched the browser or something else caused the bookmarks menu to be lost. Redoing all that stuff after it got wiped was much more of a chore than doing it initially.

But this would be my takeaway here: If software developers want their users to manage their own ordering schemes or use customization features, make it easy and convenient to do - and for the love of everything, threat the resulting bookmark trees, settings files etc as critical data that has to be synced, backed-up etc. If it has to be redone after a few updates or a crash, then no one will use it.

(Incidentally, tech companies have absolutely no problem "educating" their users if the behavior is something they care about. I'm pretty sure, no product or marketing person ever said "oh well, it sure would be nice if all our users made accounts or installed our app, but they're far too lazy for that, so that's a nonstarter...")

By that logic, most of HN is "empty calories". Then why are you here?

> If matters had transpired differently, All New World of Lemmings would have been the banner title of a new sub-series, with different subtitles for each of its entries. As it was, Psygnosis just dumped the would-be first entry of the sub-series out there on its own, sans subtitle, and abandoned the other nine tribes to their fate.

This reads to me as if Psygnosis only accepted the first of the proposed four games and made it into a standalone title - which then turned out bad anyway.

(Not sure how DMA's four games match with Pygnosis' contract for one more game. Maybe that was an attempt by DMA to extend the contract?)

> Not long into development, a feeling took hold in Dundee that they were all just done with Lemmings, ready to come out from under the shadow of the casual monster they had accidentally created and return to making the games they wanted to make. As it happened, this was to be the last game of a six-game contract they had signed with Psygnosis back in the 1980s. Just get ‘er done became the watchword, so that they could take their Lemmings loot and do something more exciting with it.

Sounds like the publisher was the sane one here for a change, while the studio was seriously trying to bite off more than it could chew.

What made them think it was a good idea to commit to four Lemmings games in advance, lock themselves into a rigid storytelling structure, extend the story into a shared universe and coordinate with a TV show - if they all were actually fed up with the franchise at that point?

Thank Psygnosis for rejecting this and enabling the timeline in which we got GTA instead.