baq

27 ms·

691.
by baq 7mo ago
Arguably it’s a very useful property of prediction markets as long as everyone is aware that’s the case.
692.
by baq 7mo ago
Interesting, I made me think when was the last time I wanted to look at the App Store and it was… never? I mean I have to to install something, but I never go there just to window shop. (My kids do, but that’s because the games have video
693.
by baq 7mo ago
I assume Mythos, if ever released to the wide public at all, will only be available in the Claude cloud harness. (Not counting special enterprise and government contracts naturally.)
694.
by baq 7mo ago
> why would you even want to use their services in the future? Uhhh my base case is you will be forced to or just be forgotten, not unlike not having a cell phone or a bank account.
695.
by baq 7mo ago
this would be a worse crisis than we've just had; it'd put China (if not all of Asia) directly against the USA and would put Australia in a very peculiar spot.
696.
by baq 7mo ago
Everyone, once AI diffuses enough. You’ll be unhireable if you don’t use AI in a year or two.
697.
by baq 7mo ago
good luck getting mythos/spud quality models open
698.
by baq 7mo ago
the problem is you're still a year behind with this approach and it isn't at all clear locally hosted models can keep the gap. need more turboquant-like algorithmic boosts for this to happen.
699.
by baq 7mo ago
$2M/ship is $100B/year at pre-war crossing rates.
700.
by baq 7mo ago
> You can't stop accepting new customers unless you're fine with killing your potential future customer base. That's a ridiculous suggestion. what? they already have, they aren't releasing mythos except to a limited p
701.
by baq 7mo ago
you could arguably ditch the productivity suite and a few other 'essential' subscriptions to make room for this one, except the price point will get enshittified to hell in the coming months and years.
702.
by baq 7mo ago
they can adjust prices or you can adjust your expectations. ...and let's be realistic, it'll be both.
703.
by baq 7mo ago
Anecdotally me and my colleagues haven’t written a substantial line of code since January and this isn’t a mag7; I would be very surprised if mag7 were writing anything by hand unless it’s a custom DSL.
704.
by baq 7mo ago
— someone in 1918, probably
705.
by baq 7mo ago
> Relative to what? to WW1 and WW2.
706.
by baq 7mo ago
in that case in particular it led to 80 years of relatively calm geopolitics kinetically, all things considered. I'm not sure I want to live through an AI cold war, but it sure seems I don't get to choose.
707.
by baq 7mo ago
don't confuse bayesian skepticism with plain old contrarian bias. a true bayesian updates their priors, I'd say this is an appropriate time to do so. also don't confuse what they sell with what they have internally.
708.
by baq 7mo ago
> It's messed up that Anthropic simultaneously claims to be a public benefit copro and is also picking who gets to benefit from their newly enhanced cybersecurity capabilities. It means that the economic benefit is going to the exis
709.
by baq 7mo ago
...and codex is at least 10x better than Claude. I don't even bother starting a new session when working on a feature, a single compaction is basically unnoticeable. You have to compact several times to start needing to remind the mode
710.
by baq 7mo ago
Not sure how Claude and CC has become the defacto best model given gpt 5.3 codex and 5.4 exist. This space moves so fast you should be testing your workflows on different models at least once every quarter, prudently once a month.
711.
by baq 7mo ago
Persian gulf is everyone’s business.
712.
by baq 7mo ago
The longer I live, the more secrets coming out I see, the less surprised I am with every next one.
713.
by baq 7mo ago
Absolutely not the case. They're complementary.
714.
by baq 7mo ago
This is the way. Eg. IME Gemini is really damn good at sql.
715.
by baq 7mo ago
I’m one of those ‘us’, Claude’s outputs require significant review and iteration effort (to put it bluntly they get destroyed by gpt and Gemini). I’m basically using sonnet to do code search and write up since it is a better (more human-lik
716.
by baq 7mo ago
Do you know of evals with default Claude vs caveman Claude vs politician Claude solving the same tasks? Hypothesis is plausible, but I wouldn’t take it for granted
717.
by baq 7mo ago
cursor should be advertising multi-model adversarial reviews, I do this all the time and let me tell you things that slip through the cracks when opus or gpt write code that gemini catches are downright scary, on the backend anyway.
718.
by baq 7mo ago
X was the primary source, it's been since reported all over the news.
719.
by baq 7mo ago
I don't think they'd be welcomed at all is the point...
720.
by baq 7mo ago
Everybody is acts so surprised as if nobody (around here of all places!) read the sama tweet in which he was hiring the Head of Preparedness... in December. https://xcancel.com/sama/status/2004939524216910323

More ›