Leonhard

@cremieuxrecueil really surprised about no less agreeable

How do the personalities of rich people compare to the personalities of everyone else? They tend to be less risk averse, less neurotic, more open to new experiences, more extraverted, a bit more conscientious, and no more or less agreeable.

@BitPaine good on you for owning up to this

there were signs

now that iran can sell oil again, wouldn't they want to sell it for bitcoin?

@JavierBlas do i understand right that they can sell in usd, but don't have to?

It may only last 60 days (I doubt it; likely would become permanent), but the US Treasury’s waiver is effectively rolling back 40-year plus of American oil sanctions against Iran. Even US refiners are allowed to import Iranian oil and pay for it in greenback!!!

@hosseeb great read, but wouldn't the natural evolution for the big co's to drift towards open source models more? i can see how they may not want to give up on the frontier for coding and research, but weaker models will be fine for most of other use cases

@JustinGorya I can't see anything about free GLM 5.2 or Kimi 2.7 on their website, this sounds a little too good to be true. What's the source for this?

👀Remember Devin? The $500 SWE? you don't need $500 anymore.. don't be afraid It’s actually heavily underrated! A $20 subscription gives you: - Devin Cloud access a fully autonomous SWE - FREE unlimited access to GLM-5.2 - FREE unlimited access to Kimi-2.7 - Access to all SOTA models - Devin CLI for your local workflow - Devin Desktop if you prefer working with an IDE - ACP connection to almost all providers in Devin Desktop If you used Windsurf Cascade and are now afraid of Devin Desktop, give it a second try and DROP Cascade. ONLY use Devin Local. Trust me Devin is really good and token-efficient. Cognition and Devin are cooking.

@reallyoptimized this whole "debate" was an interesting insight into how the medical field seems to have absolutely zero awareness of feedback loops in software

These radiologists are in for a rude awakening when the future inevitably arrives. Midjourney is ultrasound. Ultrasound is *already* reducing the need for biopsy. It's only going to get better and more robust - rapidly. The next stage will be treating the cancer with the ultrasound as well, which is also *already* happening (fully FDA approved) and will only expand. Directionally, this means we are rapidly progressing to a future where the machine could perform the imaging, analyze it effectively and treat it with little to no human intervention at all. How soon we get there is debatable, but with exponential progress, it's likely sooner than almost anyone can imagine.

Well you really can’t build an AI lab in Europe.

turns out the company behind GLM 5.2 is public and it's currently trading at ~$119B on ~$92M revenue. 1,289x multiple just so you know how crazy this is. anthropic with $47B rev would be worth over $60 TRILLION

@plunderstruck @Doctor_JDawg @thsottiaux same here, running multi-day loops plus ongoing work

@Doctor_JDawg @thsottiaux Bro what you working on how 😂

Now that you can bank usage resets in Codex. Are you a hoarder or do you use them without breaking a sweat? How do you think about them?

@thsottiaux i'd use them more aggressively if they didn't result in extending the countdown to the next reset

@sonikudzu @PrinceVogel @slatestarcodex @drethelin thanks, this is well-articulated. i agree that there could plausibly be a temp increase in demand on healthcare infrastructure, but a) this is by no means certain b) i doubt this is the real reason for the visceral reaction we've seen

okay, responding to two aspects here: 1) “people are too stupid to make their own decisions” is an oversimplification of a real principle in public health. the TLDR is it can be wild expensive not to consider this when making ~any rules and recommendations, but that doesn’t mean it should block everyone from doing anything 2) simplest example of a real problem it’d cause: doctors largely do not get to decide case by case what to do about things. there’s a concept called standard of care, and in order to deviate from it, they have to be able to justify it with evidence. so if there is a huge uptick in early findings, that means a huge uptick monitoring(at least), which requires appointments and tests. we already have a dearth of available appointments with pcps and pretty much every relevant specialist, so a major uptick in appointments/testing bc guidelines say you must check on x growth in x months strains an already strained system, delaying care for people who are actually sick. and this doesn’t even touch on the 2nd and 3rd order effects! i’m not saying this can’t be worked on, but “a solution likely exists” doesn’t negate the real problem that would occur. you have to try to account for serious patient volume changes or everybody suffers

What???

@elonmusk France, Germany, UK are no longer free societies.

Wow

from what i'm seeing the md's critique of scanner from midjourney seems to boil down to "people are too stupid to make their own decisions" kinda reminds me of catholic church *really* disliking the idea of translating the bible

@thsottiaux gib $500 plan, switching back and forth between account is super annoying

What should we improve in the Codex app. What's not delightful?

@sonikudzu @PrinceVogel @slatestarcodex @drethelin can you give examples of those real problems? i have not read every post, but so far what i am seeing boils down to "people are too stupid to make their own decisions", kinda reminds me of catholic church *really* disliking the idea of translating the bible

no, i think a lot of people aren’t getting that the healthcare system is a bureaucracy riddled with algorithms and dependencies that can’t just be changed all at once tomorrow even if everyone involved wanted that(and they do not). the things y’all are handwaving away are real problems that will arise if certain assumptions change, and some people quite reasonably want to have thought through how to handle it before the problems get out of control. there are people who are just reacting, but there are real concerns that will need to be addressed

@PrinceVogel @sonikudzu @slatestarcodex @drethelin i'm starting to think that maybe they don't get it because they just don't understand data loops and the only framework they can work from is the reality where their meat brains is the limiting factor on information processing

@sonikudzu @slatestarcodex @drethelin Yeah so how about we find out with better technology which we use to collect more data for God's sakes We know *right now* that most of the little guys resolve Just *tell people that* and keep an eye on it

@PrinceVogel @Laserpig_Utopia especially that there isn't that much left of it post covid

@Laserpig_Utopia Doing so is destroying their public reputation, which is the very thing they regard as a great evil! x.com/i/status/20684…

Every single doctor is hazed through a gruelling, expensive, long and ultimately mediocre training system. Almost all of them will hold onto the social value of that credential until their fingers bleed. Their identity is at stake, they can concede no ground

@donaldtusk have some dignity, if not for yourself, then for the people you're supposed to represent.

Brnięcie w konflikt polityków w Polsce i w Ukrainie to strategiczny błąd, na którym stracą obie strony: biznesowo, geopolitycznie i reputacyjnie. A w polityce, jak wiadomo, błąd jest gorszy od zbrodni. W rozmowach z moimi europejskimi partnerami staram się minimalizować straty i obniżać napięcie. Nie jest to łatwe zadanie.

@EastlondonDev Step 1: dissolve EU

I am worried about the future in Europe right now. There are no foundation AI labs here. There are no GPU fabs here (despite having ASML) and no plans to build or upgrade any foundry with sub 7nm capabilities (despite EU CHIPS act). There is no energy policy for the power AI data centers need. There are pockets of excellence but the strategic foundations are totally missing. Doing any work at the technological frontier of tech from manufacturing to robotics to AI is dependent on Asia and the USA (try making an autonomous robot without chips and GPUs from Asia or training one without AI models and compute from USA). Why is Europe so out of the loop (building 12-28nm fabs in Dresden, WTF are we gonna do with these?)? How do we get on track?