Sabitlenmiş Tweet
The glass isn't half-full nor half-empty. It's twice as big as it needs to be.
English
Steven Bower
930 posts
Steven Bower
@bowerblu
Originally from Mars, just waiting for a ride home.
Knowhere Katılım Aralık 2013
517 Takip Edilen235 Takipçiler
Steven Bower retweetledi
Attackers have frontier AI. Defenders need a frontier AI ecosystem—the best open and closed models, force-multiplied by a global community.
During the Hugging Face incident, closed AI blocked essential forensics. An open-weight frontier model helped contain the intrusion.
That’s why we created the Open Secure AI Alliance.
NVIDIA@nvidia
AI security advances when the industry builds in the open, together. We're introducing the Open Secure AI Alliance with industry leaders to develop new techniques and tools to safeguard software and agents. By sharing models, tooling and research in the open, we can broaden the community of defenders. Learn more about the founding members’ contributions: nvda.ws/4pD8Fc5
English
@br1ttany I’ve got a few really good ones, but not the easiest to find, and usually small independent publishers.
English
Can confirm the content of children's books has degraded along with the artwork, and it's painful to imagine adults reading them repeatedly. I've just about given up on buying contemporary books for my friends' kids
Boze Herrington, Library Owl 😴🧙♀️@SketchesbyBoze
Call me old-fashioned but I think kids were a million times better off when they grew up reading books with these sorts of illustrations in them.
English
We will recall this time in history as the Turbulent Twenties.
Play our hand right, will be followed by the Triumphant Thirties.
The future will be wilder than sci-fi could even imagine.
English
Genuinely impressive... fully AI feature length films coming faster than anyone expects.
Heavy Pulp@heavypulp
Dialogue scene from Homer's Odyssey built on Grok Imagine. Thread on how this was built below.
English
@elonmusk When models to get good at product design and manufacturing, the latent productivity unlock will be massive.
English
Everyone still stuck on software engineering benchmarks.
Prediction: within a year, all models will be so good at software we will stop caring about measuring it, and start focusing on other surprising capabilities.
0xMarioNawfal@RoundtableSpace
Gemini 3.5 Pro benchmark leak just dropped and the numbers are turning heads. > Reportedly outperforming Claude Fable 5 and GPT-5.6 in internal evals > Significant zero-shot performance improvements over 3.1 Pro > Currently in private validation and testing - Public rollout preparations appear to be underway - July 17 is the targeted launch date If the leaked numbers hold, this isn't Google catching up ,it's taking the lead. Fable 5 just became the model to beat. Gemini 3.5 Pro might already have beaten it.
English
SpaceX’s massive corpus of world-class engineering data (excluding material blocked by ITAR) will be added during supplemental training of the 2T run.
This will dramatically improve Grok’s engineering capabilities.
Cory Robin@flycory
For the past few wks since Grok Build launch, we've run SxS testing of complex Codex & Grok Build tasks. Eliminated Claude for needy intern BS. Grok Build wins more trust w/ higher accuracy & less hand-holding. Codex remains solid; each has unique strengths. For heavy math/science/data in aerospace, we're normalizing on Grok Build as of Jul 20. Exciting yet nerve-racking shift after 6mo on Codex. Qs on process/A/B testing welcome. Congratulations to @SpaceXAI and the Build TEAM!
English
@PalmerLuckey I for one welcome “Carmack” as the new SWE benchmark. How many Carmack’s is Fable 5?
English
Everyone who thinks AI slop will ruin code efficiency/performance is going to be so surprised when everything is absurdly well-optimized John Carmack style machine code.
English
@beffjezos Turns out big models aren’t that hard once you have the right inputs and enough compute turned online. Expect major “labs” to keep one-upping each other every few months for next few years. Capabilities will converge and then asymptote quickly.
English
The good news is: all the doomer predictions are always wrong. Without fail. Every. Single. Time.
Stop listening to them.
Michael Shellenberger@shellenberger
The CEO of Anthropic, the AI company behind the popular Claude chatbot, believes "it cannot possibly be more than a few years before AI is better than humans at essentially everything.” Indeed, the company's near-trillion dollar valuation rests on a plan to destroy millions of high-paying jobs. Not to worry, says the person at Anthropic in charge of "research partnerships with the world’s wisdom traditions," since all of that destruction will free people up for things like "grounds maintenance," "food and serving," and "personal care." This is dystopian. Those are the lowest-wage jobs in America. Millions of more people in the service sector will only further drive down wages. Anthropic's head of researching wisdom traditions tries to spin this by saying, "another word for grounds maintenance is gardening. And another word for food and, and service is hospitality." Anthropic says we should think of this as "The Great Turning," which is no different from the World Economic Forum's equally dystopian "Great Reset." Part of what's behind Anthropic's efforts to humanize and even spiritualize AI is the need to convince investors that its technology will destroy millions of high-paying jobs. But another part is a deeply degrading and dehumanizing view of people. x.com/shellenberger/…
English
@shellenberger The good news is: all the doomer predictions are always wrong
English
The CEO of Anthropic, the AI company behind the popular Claude chatbot, believes "it cannot possibly be more than a few years before AI is better than humans at essentially everything.”
Indeed, the company's near-trillion dollar valuation rests on a plan to destroy millions of high-paying jobs. Not to worry, says the person at Anthropic in charge of "research partnerships with the world’s wisdom traditions," since all of that destruction will free people up for things like "grounds maintenance," "food and serving," and "personal care."
This is dystopian. Those are the lowest-wage jobs in America. Millions of more people in the service sector will only further drive down wages. Anthropic's head of researching wisdom traditions tries to spin this by saying, "another word for grounds maintenance is gardening. And another word for food and, and service is hospitality."
Anthropic says we should think of this as "The Great Turning," which is no different from the World Economic Forum's equally dystopian "Great Reset."
Part of what's behind Anthropic's efforts to humanize and even spiritualize AI is the need to convince investors that its technology will destroy millions of high-paying jobs. But another part is a deeply degrading and dehumanizing view of people.
x.com/shellenberger/…
Michael Shellenberger@shellenberger
English
UBI politicians need to chill out.
We need to reach that level of increased productivity first before we even discuss UBI.
Anything before that is financial suicide
Andrew Yang🧢⬆️🇺🇸@AndrewYang
👀
English
@DigitalBarter @TheGameVerse "It's hard to sense care, but you do sense carelessness" (paraphrasing Jony Ive)
English
@TheGameVerse Only people who are mediocre in their real life won’t appreciate it. It’s not about you care or not. Good design often means you don’t realize someone has worked on that problem to make your experience smooth. Only bad design makes us realize there was a problem.
English
@chamath Team should be actively looking for which AI models hit token efficiency vs productivity nexus. Too many stick with whatever model family they started with and tokenmax latest version. Token efficiency is the new vital metric after cost. (Grok 4.5 vs Opus 4.8, as an example).
English
8090 works on production systems for large, often regulated, enterprises.
Vibing isn’t tolerated because these are the systems that run western society - banking, power, healthcare, insurance etc.
Over the last few quarters, the gains that we got from using frontier models inside of our Software Factory on these systems started to shrink but the costs kept doubling. This makes sense I guess, as in hindsight, we were initially asking the model to do mostly light work (generate basic PRs) and now we were asking it to do more complex work (mitigate dependencies across systems).
Unless you grow context massively, be willing to run many A/B tests and iterate massively (ie use massively more tokens) complex tasks stay roughly unfinished by the model and requires the engineer to largely act alone.
In other words, we find the last 5% (ie where a model is truly equivalent to a reasonable engineer) extremely difficult to achieve and extremely expensive to such a degree that the fully loaded cost of the model + the engineer will not pay for itself.
So I asked our CTO to start thinking about other ways. We need our engineers to have access to the best tools BUT we also need to educate them to think even more for themselves - not less - in this last mile.
At the same time, we need to find solutions that decrease our token costs by 90% - especially because these bleeding edge tokens are not nearly as cost effective as the tokens before it and are creating a big OpEx bill for us.
I wonder how many engineers, in all orgs, are running amok right now by using the latest frontier models as a kind of slot machine. Increasingly turning their mind off, largely keeping productivity flat while their CEO and CFO deals with a massive token bill?
My advice to you is that when you encounter this last 5% of very hard technical challenges in getting a complex system into production, be circumspect.
The challenge of the last 5% is actually getting harder - especially as hundreds and thousands of code generation model runs run amok adding all kinds of random cruft into codebases that eventually need to be rationalized.
dnap@dnapway
Chamath reveals his company's AI token costs are doubling every 45 days but productivity is only up 5% "I sat down with my CTO today, I said how are we doing on token spend. And he said the most incredible thing, he said right now, our token costs are doubling every 45 days. I said well what is the downstream productivity? And he said maybe 5% max." "So my costs are doubling every 45 days, my upside is essentially flat. He said honestly, what we're finding out is that you need to use a lot more tokens to get to this next iteration of improvement because we've effectively already asymptoted." "We're going to take a step back and try to figure out what to do. I don't know how many other companies will actually go through this reckoning now, but the point is everybody in the next three or four years will for sure go through it." "I suspect that if you can get out now, you should get out now before all of that starts to seep into the water table. Because I think that's probably what allows you to get out at a huge price and raise a huge amount of money."
English
@jasonfried You’re not alone. Right now I can’t imagine buying any other car.
English
The current Y is the best overall car I've ever owned, and I've owned plenty across the spectrum. Nothing's even close.
DHH@dhh
I've been a Tesla convert for years. Model X, Model S Plaid, Cyberbeast. But the Model Y we just got still blew me away! It feels so tight, so well built, so right. Unbelievably good car and good value. It sending an Aston Martin DBX 707 and Taycan Turbo S out of the fleet!
English
@richgel999 by then LLMs wont care about our pointless abstraction layers anyway and will just write the raw machine code
English
be the time C++35 arrives nearly all code will be written by LLM's and I probably won't care how ugly C++ can be
pastaya@realpastaya
Multiply a variable amount of numbers in C++35 (the fast, safe and pure way).
English
@levelsio LG has no such limits, but still requires cloud to control AC from Home Assistsnt. There’s a few AC models people have reverse engineered for local control. (Which should be legally required imo. Yes, I will die on this hill)
English
Never ever buy a Daikin AC!
They limit how many times you can change them to just 150 times per day!
We have 7 ACs in the house so that's a limit of 21 instructions per day 😂 completely arbitrary btw, makes no sense, just classic Japanese software design
(I didn't buy it, it came with the house, and I've actively been ripping them out and replacing them with Mitsubishi)
Other good brands for ACs are LG, Hisense and Haier, can recommend!
@levelsio@levelsio
Today my Daikin AC required me to accepts its new terms and conditions to continue using it
English
If models themselves become commodities, the model maker valuations will burst quickly. That’s why private model makers are pivoting to hardware. The physical layer is where actual value is.
All this fear mongering about capabilities will be meaningless in a few years.
Chamath Palihapitiya@chamath
We’ve had every major open source models provider and CSP reach out to 8090 since the post below. We will run a massive recursive optimization with the major open weight models (Chinese) AND American Open Source - starting with Nvidia’s Nemotron - and will publish the results. I think you will see, unambiguously, that a 3rd party control plane (like 8090’s Software Factory) + open weight/open source + leading CSP is much cheaper and much more strategically protective than using a closed model directly.
English
It’s such a pedestrian use case, but I find LLMs are incredibly good at doing product research and review analyses; finding me best value/price for any product I am looking the buy.
The amount of time this saves me is so much more than I expected.
English