You would need a way to walk through a good chunk of realistic gameplay without user input. Good luck with that.
The closest thing to this right now is the performance monitoring that Valve does for Steam Deck compatibility — which isn’t headless, it’s just crowdsourced.
That’s excellent. I think the initial momentum towards a DSL is basically equivalent to https://en.wikipedia.org/wiki/Inner-platform_effect
Reminds me of: https://timharford.com/2024/07/8801/
He’s been jockeying for a pardon for a long time now, but he made the mistake of cozying up to Biden during the election.
Not even staging an interview to talk about how he finally saw the light and came around on MAGA can save him now.
A spec that’s more than a single page rarely survives initial contact with reality.
The trouble with “easy” languages is that they allow you to write incorrect code without much pushback.
The trouble with “serious” languages is that they push back even when you’re writing correct code.
There’s plenty to learn from either route. Just pick something that piques your interest and dive in.
Yep. And that’s exactly why we need to keep electing self-proclaimed socialists.
That’s quite a cherry-picked quote, and to connect it to Bernie and AOC is to completely disregard the rest of the text.
Rosa draws a distinction between being a legislator in a parliament – where you’re able to battle the state within its own arena, and expose its contradictions – and being an executive in a cabinet – where your job is to manage crises and therefore hide contradictions.
But coming from a minister, social reforms can’t have the character of the proletarian class, but solely the character of the bourgeois class, for the minister, by the post he occupies, attaches himself to that class by all the functions of a bourgeois, militarist government. While in parliament, or on the municipal council, we obtain useful reforms by combating the bourgeois government, while occupying a ministerial post we arrive at the same reforms by supporting the bourgeois state.
She also explicitly endorses political engagement as a means of class struggle:
Personally, in this great gathering of the different socialist organizations in the free play of the daily political struggle, we don’t fear the least danger for the doctrine of Marx and the principles of democratic socialism, in as much as they have already taken root in France. There is no better school for socialist democracy than the great and living class struggle freed from abstract clichés. The materialist conception of history doesn’t allow us to believe in the development of a living popular movement begotten of abstract formulas; on the contrary, it’s on the material base of a great and strong class struggle, embracing all of the proletariat, that a clear conception of theory and principles will be erected.
And not for nothing, the Communist Manifesto does too:
In short, the Communists everywhere support every revolutionary movement against the existing social and political order of things.
In all these movements, they bring to the front, as the leading question in each, the property question, no matter what its degree of development at the time.
Finally, they labour everywhere for the union and agreement of the democratic parties of all countries.
The broader theme of Rosa’s text is to not accidentally align yourself with liberals in a cross-class coalition by focusing only on the political question of the day. I think we’re in danger of doing that with memes just as much as self-proclaimed socialist politicians are in danger of doing it by entering government.
It’s a matter of feedback loops.
It’s the same problem as when you divide teams by front-end/back-end, or implementation vs testing, or features vs platforming.
When you don’t have to feel the pain of your decisions, you’re going to make bad decisions.
There’s an episode of the You Are Not So Smart podcast where the guest (either was? or just interviewed?) the psychologist in charge of interviewing astronauts for suitable personality traits. He had a ton of great insights, but the thing that stuck with me is that there are basically three kinds of conversations:
They’re all important, but having the right one at the right time is the key.
My original comment was flippant, but you’re fightin for your life with earnesty in these replies, so I’m gonna provide what little IRL experience I have in this so far.
Spec driven development is better than ephemeral prompting in the short term while building up… but for maintenance, it works against you.
Once the code has been generated from the spec, the two inevitably drift apart.
Even for humans, this is a problem. Stale docs can waste time and mislead developers, so the best spec is one that is executable to confirm that it still matches the implementation.
But for agents, it’s especially important, because they have a harder time detecting stale docs and disregarding them, and also because LLMs corrupt documents over time, so they will invariably cause this problem.
So the best spec turns out to be tests. Which means your spec is gonna be code, not natural language. (Djikstra has some insight there.)
Which means you need it to be easy to write tests as a human. Which means you need to aggressively refactor. Which agents are not great at doing.
But even if they were, the dirty secret about refactoring is that it’s heavily dependent on having a good taxonomy for your subject matter, which is a people problem and not a coding problem.
My own sense of this landscape is that AI is effective in two radically different scenarios:
I took a course where the capstone was to achieve Ralph Loop Nirvana, to show the suits that I’m willing to play along. But I’m unconvinced. If you let AI consume input that is mostly AI-generated, it seems to inevitably deteriorate.
There’s no avoiding it: you simply must get your hands dirty in order to keep things organized, and that’s at odds with any of these “your new job is to be a manager, or SME, or PM, or whatever” tactics.
—
Adding another thought:
People really underappreciate friction.
We spent a long time living in a world where friction was unavoidable. Then with digital automation, we got a taste of what a very low friction world would look like. And we got used to the idea that reducing friction is an absolute good.
But some tasks need friction in order to succeed. And since friction was never optional before, we never noticed that potential failure mode.
Because of this, the sales pitch that AI will remove all friction is fundamentally flawed. But that doesn’t mean AI has nothing to offer for those “friction-required” kinds of tasks.
AI is actually great at grilling you about ambiguities, contradictions, and dependencies. You can use it to deliberately apply friction in the right places.
The problem is… most people don’t use it that way. They use it as something that just implements and doesn’t second-guess, and they miss out on the cases where it’s useful to have it second-guess and not implement.
Spec driven development is a scam.
Is there a word for AI slop + AI psychosis?
Can I still eat a dick?
They wrecked the economy so hard that game consoles are appreciating in value.
I’ve been thinking for a while, that the thing AI excels at is not completing work (obviously) but making incomplete work someone else’s problem.
In that way, it really is the perfect embodiment of capitalism in action. Less efficient overall than actually successfully doing the thing, but more profitable for a single atomized individual acting in their self-interest.
I believe that they use it all day. I believe that they say they find it useful.
I also believe that their bosses gave them a productivity slot machine and told them if they don’t play it they’re fired.
So some of them like it for bad reasons, and some of them have to pretend to like it.
Lmao no they’re not. Suits love to pretend they are, but it’s been a lie ever since the days of “scientific management”. Falling for that lie is why Deming and lean manufacturing were such a revolution.
People who don’t history something something
I hate to be skeptical of someone who lost a child in a traumatic way, but this does read like AI psychosis.