Prakash (Ate-a-Pi)
@8teAPi
FOLLOW ME for AI commentary; tech optimist, future shocked self-aware neuron, once fooled by superconductors;
What I hope to teach my kids? In an AGI future, what should they learn? A) Risk Perception and Appetite > no AGI can solve this > every decision will come down to several choices with some probability of success > you must assign your own utility function to weigh those…
Dropping off a cliff. The entire Chinese system is processing a 25% drop in population. When these kids hid child bearing age that will compound again.
Financial Times: “The number of Chinese kindergartens has fallen by a quarter in four years.” How is that even possible? East Asians simply have not been able to adapt to the smartphone era. Something is deeply wrong in all their societies.
Facial recognition glasses - All the societally unpleasant stuff big tech was keeping quiet about, killing project demos internally, are getting build by startups. To some extent this is what democratization looks like.
It should be reversible and common. You should be able to grow a pair of wings to go hang gliding in the Andes, rest overnight and get on your Starship back to Denver as a normal human
We're going to be having some fun conversations about the ethics of gene modification over the next few decades. Like "is this eugenics?" What about deafness or blindness? Tay-Sachs? There's no clear line.
If AGI is coming, go have a meal at a restaurant with human staff. There will come a time when only billionaires can afford human labor to serve a meal. Wait staff are going to go the way of horses. One day we will wonder at how much human labor was wasted on tasks robots can do
Stanford actually breaking the chain here. The RAs normally select the next year’s RAs through nepotism which is how the co-ops maintained their character. Replacing with generic RAs and then having them select the next generation is the end of it.
Stanford has suspended two student residential co-ops for a year for Title VI violations. One co-op was suspended for targeting Jewish students and the other for exclusionary behavior toward white and male students.
Is this Stargate? Is this actually getting built? And how in the world are they building power plants in 200 days ?
OpenAI Abilene Texas Construction Leak Crusoe Energy is the construction partner for Oracle/OpenAI - 5,000 workers in Abilene right now - Abilene chosen because massive oversupply of wind energy in Texas causing prices to go negative - Phase 1 took 300 days, Phase 2 will…
I’d expect these to be banned or heavily tariffed in the US. Let’s see how long it takes.
🚀🔥 The Unitree R1 just landed — and it's an absolute monster. Built for locomotion supremacy, it packs 26 DOFs into just 25kg of pure speed and agility — all for ~$6K, nearly 10× cheaper than the G1 🤯 Looks? Straight out of a sci-fi film, just like a space explorer 🧑🚀 We’re…
Nope. It’s because things which are not under human control fall mentally into a category of background. For eg climate, religion, financial markets. Each of those is arguably an information system that affects our lives. But we ignore them because they don’t seem to have human…
I think I finally figured out why some people are weirdly confident AGI will "obviously" stay under human control indefinitely: They've been programmed by 70 years of science fiction where intelligent machines serve us without complaint. Data takes orders from Picard. Star Wars…
Broadcom, now a $1 trillion market cap Nvidia competitor, started off as the private equity buyout of former Hewlett Packard division Agilent’s Semiconductor group. Silverlake and KKR bought the division, hired Hock Tan to run it. He took it public, gave them their exit and then…
So there's a persistent idea floating around that goes like "Private equity firms buy up good companies and then run them into the ground, and this is a huge and persistent problem". This confuses me: why would they buy companies just to ruin them? Wouldn't they lose money? PE…
Claim is that they used an agentic system to do architecture search and hit SOTA in some benchmarks.
This paper makes a bold claim! AlphaGo Moment for Model Architecture Discovery The researchers introduce ASI-Arch, the first Artificial Superintelligence for AI Research (ASI4AI), enabling fully automated neural architecture innovation. No human-designed search space. No human…
If the stablecoin bill works, they could do this. Let’s see.
6/9 It can only do so in one of two ways. Either it forces a change in the structure of the global economy that eliminates the abilities of individual countries to control their trade and capital accounts in order to externalize...
GPT-5 in the wild, it begins
BREAKING 🚨: Microsoft is preparing Copilot for GPT-5 release with a new Smart Mode! "Smart Mode uses GPT-5 to think quickly or deeply" Was GPT-5 release delayed b/c of that? 👀
Excel Killer: tryshortcut walkthrough x.com/i/broadcasts/1…
Within 3 years of launch, ChatGPT has completely killed the use of all high school and much of college education This is a “gray rhino”, a fact obv true but so large that it is indigestible. An enormous risk in plain sight that everyone ignores by social convention.
Incredibly damning article about 1 of the Top 3 MBA Programs in the country. I’ll try to get some key excerpts here but I would suggest reading the whole article.
Ground drones coming next
DJI is entering the robot vacuum space. My DJI friends who have been using beta builds in their homes say it's the best on the market, by far. It's going to be pricey.
AI interviewers to waste more job candidate hours for the win.
Ready for an AI interview? That person interviewing you just might be an AI. Seattle startup Humanly launches real time AI avatar interviewers modelled on your company's staff.
Vogue has fallen
boycott Vogue NOW! they’re pushing out models for AI, devaluing hard work and art of real people.
Stablecoins, force/allow the rest of the world direct deposit to US dollars. Lower US interest rates by crowding in to money market funds. No KYC/AML. A financial solution for a financial economy.
do boomers realize they're mostly still gonna be alive when we soft default in like 5-15 years thinking about this a lot lately what's their strategy here
30% of Humanity’s Last Exam biochem questions have incorrect answers. You’ll know your model is smart when it yeets the entire eval telling you it’s wrong.
HLE has recently become the benchmark to beat for frontier agents. We @FutureHouseSF took a closer look at the chem and bio questions and found about 30% of them are likely invalid based on our analysis and third-party PhD evaluations. 1/7