It is amazing that human language turned out to be so unreasonably effective at modeling the universe that training a system to predict what comes next in a sentence would get us this far.
Something I think people fail to understand is that as soon as there is 1 cheap / smaller Chinese open weights model roughly on par with Opus 5.5, it is game over for Anthropic. 5.5 is, for better or worse, good enough for almost every task. 🧵
Weet niet of het puur en alleen belastinggeld is? De vraag is ook wat er gebeurt wanneer je een plafond zou instellen. Denk je dat de beste specialisten niet op een andere manier gaan verdienen wat ze waard zijn? Prive-klinieken, buitenland, etc? Voor mij is het een fijn idee dat de beste specialisten geen reden hebben om buiten het gereguleerde systeem te gaan werken.
Seaport just won the internet, and it's not even close:
"Anthropic could develop AI systems which become fully autonomous and wipe out humanity which may lead to a serious decline in earnings."
Update: my dear friend has now seen my thoughts. And because he is genuinely one of the most thoughtful people I know, I must add his response :
‘Fwiw I would still say 1) adoption revenues are needed at the very least as a bridge to support getting to the infinity machine (especially if it takes 5 years rather than 1) and 2) an infinity machine still isn't the same as financial return and also far from guaranteed the labs are the ones in the supply chain to capture the value. But I do agree your scenario is the one that makes the capex approach most clearly stack up.’
We also both agreed that if compute is hoarded to achieve RSI (which I believe is already happening), and if the application layer is devoured as a consequence of that - then it will be ‘the worst of all worlds for Europe and a huge external economic shock.’
@brendanpmalone@matt_is_nice They will be sympathetic to incentives?
To me Paul basically says labs need to step up security because the next HF hack might have a lot more impact if we let things go as they are but with a lot more capabilities.
Dit. Omdat het wel zo snel ging heb ik meteen het idee dat er iets is dat wij nog niet weten. Maar als dat niet zo dan is het allemaal lastig te begrijpen.
last thing regarding the jensen situation. the most critical error that jensen made thus far is resigning way too quickly.
if you had a team of advisors that knew how the modern tech driven culture works, they should’ve realized that what was once boiling hot simmers & then cools off very very quickly since the volume of things that happen & shift the timeline is higher than ever before.
i am not a professional pr person (i also can’t believe how much pr ppl get paid for some basic shit). i simply like to think about how ppl fit & operate in a very new world now, esp if they have a public presence.
if you’re jensen, you should’ve simply waited it out, perhaps said something to the order of:
“nvidia is above my personal life, & this is a revolutionary moment for humanity. i am human. i am not immune to human forces. business at nvidia continues normally until a full investigation is finished at which point i will be open / honest about the realities of what is next.”
if you’ve ever done standup comedy you know that you have to sometimes buy time to hit that perfect zinger & timing, particularly in crowd work. this would’ve been a very similar principle of simply buying time. he could’ve learned a thing or two from andrew huberman too by continuing business as usual.
Ich hab grad der RDW geschrieben 📩
Erstmal ein ehrliches Danke, die Niederländer haben als erste in Europa TAD getestet und zugelassen und bleiben da echt dran 👏
Aber mit dem neuen Update kommt eine strengere Fahrerüberwachung, und da hab ich eine Frage als Papa. Mein Kind schreit plötzlich hinten, ich dreh mich kurz um und genau dann warnt mich das System oder nimmt mir die Unterstützung weg 🙈
Ausgerechnet in dem Moment, wo ich als ganz normaler Mensch Hilfe am meisten brauche, steh ich wieder alleine da. Wie soll das meine Familie sicherer machen? 🤷🏻♂️
Ich will nicht weniger Sicherheit. Ich will Systeme, die uns da helfen, wo wir nun mal Menschen sind.
Bin gespannt ob und was zurückkommt, ich halt euch auf dem Laufenden 👀
https://t.co/IvK7O58QsW