skybrian

5 ms·

61.
by skybrian 1mo ago
My cost to switch is choosing a different model from a pulldown menu.
62.
by skybrian 1mo ago
Aware of what, though? We don't actually know OpenAI's profit margins on inference. Do we need to know? It's enough to know that the market is competitive and new models with better prices are released often. This means you s
63.
by skybrian 1mo ago
Your cheap cynicism makes OpenAI's lower prices sound like a reason to avoid using them, but that's backwards! It's a better strategy to use them now while prices are low and switch later if necessary. Nobody knows who will h
64.
by skybrian 1mo ago
Well, I suppose that explains monads. It’s one thing to see an analogy and quite another to make it the basis of an API.
65.
by skybrian 1mo ago
I have a couple of skills with project-specific conventions for how to write a plan and how to write HTML-generating code. But they could probably just as well be .md files in a docs directory, linked to from the AGENTS.md file.
66.
by skybrian 1mo ago
Though it has all the signs of being AI-generated, this looks like a pretty interesting project. But by itself, it seems sort of like a game engine without a game. Have you done something interesting with it?
67.
by skybrian 1mo ago
They consider themselves to be in an arms race with all the other AI firms (including Chinese) that are not that far behind. And... are they wrong? This is why there's talk about negotiated "pacing."
68.
by skybrian 1mo ago
I have a vague understanding, enough to know that tallness isn't thinking. I think I can usually use the word correctly. (I don't think El Ajedrecista qualifies, but it was starting to play chess, so it's closer to "sm
69.
by skybrian 1mo ago
> If a model were actually capable of scheming, it would also have enough situational awareness This doesn't follow. It might not have that kind of control over its the chain-of-thought, even if in some sense it knew that would be a
70.
by skybrian 1mo ago
You're insisting on a particular definition of a vague term. It makes sense now to say that temperature is what a thermometer measures. However, before there were good thermometers, people often thought that heat and cold were differen
71.
by skybrian 1mo ago
I don't really know. Maybe a continuous build server? But just because it's listed in the catalog doesn't mean it's the one I want to use.
72.
by skybrian 1mo ago
Playing chess, writing code, finding security bugs, and proving mathematical theorems all seem fairly similar to thinking and don't seem much like being tall.
73.
by skybrian 1mo ago
"Defensive systems" can be interpreted broadly to include cybersecurity. But yes, it's an arms race. Saying it's not an arms race isn't going to make it not an arms race. Warning that it is an arms race isn't e
74.
by skybrian 1mo ago
"Smarter" is a vague term. If a bot can beat you at chess then in some sense it's "smarter" than you about chess. After repeating this feat in enough narrow domains, if you say "but it's not really smart
75.
by skybrian 1mo ago
No guarantees, but it might still be a useful signal even if it’s not causative?
76.
by skybrian 1mo ago
Perhaps this “jobs program” could be seen as preserving industrial capacity? With military contractors, for example, it’s apparently a problem when they’re used to build a stockpile and then they shut down because there are no more orders.
77.
by skybrian 1mo ago
News to me. Where did you learn about that?
78.
by skybrian 1mo ago
This pitch is somewhat interesting, but, much like Sandstorm, browsing the library of pre-built apps kind of leaves me cold. It's too much like browsing a rather tiny mobile app store. There are maybe three apps that I might be interes
79.
by skybrian 1mo ago
Shortages can't be defined out of existence this way. Markets don't magically train workers or provide security clearances. People have to do that.
80.
by skybrian 1mo ago
It's true that you need to ask good questions to get good results. If you try to understand the code yourself, you will ask better questions.
81.
by skybrian 1mo ago
A way to think of the Baumol effect in more local terms: High-productivity organizations have increased revenue and reduced expenses, almost by definition. These profits can be "spent" in various ways like charging customers less,
82.
by skybrian 1mo ago
There is a lot of noise out there from anti-AI types who never seriously tried to clean up messes with AI. If you just ask for new features and never look at the code, coding agents will make a mess. If you often ask questions about how to
83.
by skybrian 1mo ago
Why not Luna?
84.
by skybrian 1mo ago
Bluesky has live video but it's a Twitch stream. ATProto itself is about databases and ways to replicate them. You can store pictures and videos there, but you wouldn't normally use a database replication protocol to actually tran
85.
by skybrian 1mo ago
Maybe not, but I like to review the plan anyway so that I'm less surprised by what it actually did.
86.
by skybrian 1mo ago
What fatalism? OpenAI has already made security improvements and they will make more. I also think having common standards for containment might help.
87.
by skybrian 1mo ago
I think an alligator wandering down main street is worse than anything that happened at OpenAI so far, but everyone agrees it's a warning shot and could get worse. Millions of people seems, uh, much worse than that. The Ukraine war is
88.
by skybrian 1mo ago
Yes, clearly that's worse than the first time.
89.
by skybrian 1mo ago
We aren’t going to do that because intent matters. You need to control your dog and there should be penalties if you don’t, but if your dog bit someone because you didn’t control it properly, that’s not the quite the same as if you bit some
90.
by skybrian 1mo ago
I wonder if Amazon will lose enough sales to more bot-friendly vendors that they’ll notice this and do something about it?

More ›