Emerson Segura

@ErenChenAI It will probably be years and manny innovations along the way. When is Unitree going public?

Unitree CEO Wang Xingxing: Whoever builds the first Generalized Robotics Foundation Model deserves a Nobel Prize. He also said Unitree is working in that direction too, but the probability of succeeding is currently below 50%.

@ErenChenAI VLAs are far from dead, and what nivida bro described is simmilar to a VLA.. just with some key differces.. yet most of the unresolved issues with VLAs would still be unresolved in the "world model" video model he talks about ... so there is no new/diffrt leap that beats VLAs yet

Dr. Jim Fan says “VLA is dead” and that World Action Models will become the new paradigm for robotics. Predicting the next world state, fine-tuning actions, then RL for refinement. Sounds increasingly similar to how LLMs evolved. Do you think the future of robotics will be driven more by world models or teleoperation data?

@ShagunOhri Great photo

Unironic peak sf Saturday night with some v cool people

@CreviceShrapnel @TheHumanoidHub missed that 30seconds, is there a clip/timestamp?

@TheHumanoidHub Best we got was a whole 30 seconds of Omni. Everything else was boring and useless.

My very first Google I/O! Hoping to see more robotics this year.

@MarwaEldiwiny amazing

Cams act as the mechanical memory of the machine.. Jacques de Vaucanson realized that by placing hundreds of tiny, precisely shaped cams onto a single rotating cylinder , he could create a complex mechanical code I wish we had been introduced to cams differently back in college

@GaryMarcus you should be allied with people who share your views and values

seriou question: how do you handle an intellectual doppelganger who has systematically started adopting every position you have argued for for 30 years while presenting each idea as if it were his own? nothing in graduate school prepared me for this.

@mcpark @kimmonismus loll! good one

@kimmonismus Yeah… says the guy with a ai lab that can’t make models

Mustafa Suleyman says 18 months until AI automates all white-collar work. Microsoft AI CEO Mustafa Suleyman predicts "human-level performance on most professional tasks" within 18 months. Accounting, legal, marketing, project management, all fully automated. "Suleyman predicted “human-level performance on most, if not all professional tasks” being done by AI. Most tasks that involve “sitting down at a computer” will be fully automated by AI within the next year or 18 months, he said, naming accounting, legal, marketing, and even project management as vulnerable." (Fortune) Suleyman says his mission is building "superintelligence" and that creating a new AI model will soon be "like creating a podcast or writing a blog." Via Fortune

@GaryMarcus agreed, accounting does not tolerate errors, even though the math is basic, it's a zillion numbers and strategy to get a desired outcome, and llms are not close being able to do either (no numeric errors, nor "accounting/tax logic").

utter nonsense. accounting and legal will NOT be fully automated in 18 months. willing to put $100k against Suleyman on this.

A surgeon demonstrated that he could sew a balloon without popping it.

@GaryMarcus @haider1 Its incredible how big of a deal Mythoa has become in DC - anthropic are excelent at marketing

Even @haider1 sees that Mythos has been overhyped.

@NandoDF @ylecun @sainingxie @eladgil Well said

SV has the money, and the belief that they deserve the money. They also believe they are the centre of tech, and have the license to keep disrupting, ie to keep testing the law. This alone will always keep them ahead of Europe, which is overly conservative. AI ideas were born all over the world by students, professors and international teams of researchers. SV was the home of some of the most important innovations. What it didn’t invent it bought. Above all, it bought the people. Most SV engineers truly want to have a positive impact in the world, but there are also powerful narcissists, which I believe are very dangerous because their financial power is disproportionate. SV is more and more coming under the scrutiny and control of the US military and government. This may or may not be a good thing. Time will tell. Given the new world order and the growing power of AI, it is a very dangerous situation for the rest of the world. The problem is not SV, but that the rest of the world, except for China, is not prepared to invest on what it takes to be AI leaders. As @MarkJCarney said: "Middle powers must act together because if we're not at the table, we're on the menu” It is important to have a seat at the table, and for us to realise we’re all in this together. May we all look back and say: wonderful how we all came together and figured out how to build AI safely and in a way that was compassionate, universally empowering, and which created more meaning.

People at major AI labs (using internal models) 3-4 months ahead of startup silicon valley engineers SV founders/eng 3-6 months ahead of NY NY founders/eng 6-12 months ahead of rest of world Most people have no idea how fast AI shifting as 1-2 years behind SOTA "The future is here, just not equally distributed" - Robert Heinlein

@sarahookr great photo

My favorite run in sf is going across the bay bridge. It is just about long enough a run when you start from the marina to allow you to think about new ideas.

@sampullara @sarahookr Lol

@sarahookr next time call it san fran too for more rage

@itsjulespark Frontier tower

what’s a good coworking space in sf? ☕️👋

@Melt_Dem @DavidSacks are you saying you're seeing full returns on gpu+infra capex in 18 months, OR, revenue equaling capex in 18 months? (if its the fromer the implication is 18months of renting out GPU+infra, is much more revenue, like 3x than it costs to buy the gpu and datacenter),wild.

@DavidSacks even with debt service and added orchestration costs, we're trending ~18 months rn on smaller scale inference deployments. golden age of project finance, reminds me of my shale financing days.

Back-of-envelope numbers for 1 gigawatt data center: All-in Capex: ~$50 bn Enterprise revenue generated: ~$25-30 bn/year Electricity cost: $1-2 bn/year ~2 year payback. The boom is real.

@DavidSacks @eladgil how do you come up with this? Currently H100 pricing is about 20% to 40% of electricity + cooling cost... no way you make back full capex in 2y. Datacenters take years and billions (not including gpu) to make... maybe you are assuming prices for GPU inference will rise?.

@ylecun @sainingxie @eladgil SV #1 in captial deployment, sadly that matters. Also, shoutout to TORONTO, UofT for Alexnet ! (and some of Transformers paper).

@eladgil BS. Attention was born in Montréal PyTorch in NYC. AlphaGo in London AlphaFold in London ESMFold in NYC Llama 1 in Paris. Llama 2 in Paris+NYC+SV DeepSeek in Hangzhou Plus: DINO in Paris JEPA in Montréal+Paris+NYC SV is 3 mos ahead on topics SV is singularly obsessed with.

@Scobleizer @LuizaJarovsky youtu.be/_db3_5hgtAM?si…

@emerson @LuizaJarovsky My child took me to a Civil War battlefield and taught me more about the history there than the museum did. "Where did you learn that?" "I play a video game that is on the field." So don't ask me. :-)

🚨 University professors have been saying AI is completely destroying learning and that we'll soon have an AI-powered, semi-illiterate workforce. Here's a glimpse into the educational apocalypse: "Sarah, a freshman at Wilfrid Laurier University in Ontario, said she first used ChatGPT to cheat during the spring semester of her final year of high school. (...) After getting acquainted with the chatbot, Sarah used it for all her classes: Indigenous studies, law, English, and a “hippie farming class” called Green Industries. “My grades were amazing,” she said. “It changed my life.” Sarah continued to use AI when she started college this past fall. Why wouldn’t she? Rarely did she sit in class and not see other students’ laptops open to ChatGPT. Toward the end of the semester, she began to think she might be dependent on the website. She already considered herself addicted to TikTok, Instagram, Snapchat, and Reddit, where she writes under the username maybeimnotsmart. “I spend so much time on TikTok,” she said. “Hours and hours, until my eyes start hurting, which makes it hard to plan and do my schoolwork. With ChatGPT, I can write an essay in two hours that normally takes 12.” - "By November, Williams estimated that at least half of his students were using AI to write their papers. Attempts at accountability were pointless. Williams had no faith in AI detectors, and the professor teaching the class instructed him not to fail individual papers, even the clearly AI-smoothed ones. “Every time I brought it up with the professor, I got the sense he was underestimating the power of ChatGPT, and the departmental stance was, ‘Well, it’s a slippery slope, and we can’t really prove they’re using AI,’” Williams said. “I was told to grade based on what the essay would’ve gotten if it were a ‘true attempt at a paper.’ So I was grading people on their ability to use ChatGPT.” - AI in education is a serious topic, and many schools and universities are blindly jumping into the "AI-first" wave without considering short and long-term consequences. It would be great to hear more from teachers and educators to understand potential solutions. This might be a great opportunity for rethinking the education system and how students are assessed. - 👉 Link to the full article below. 👉 To learn more about AI's legal and ethical challenges, join my newsletter's 94,700+ subscribers (link below).

@Scobleizer @LuizaJarovsky "Rarely is the question asked: Is our children learning?"

This is just a total lack of creativity on behalf of educators. Years ago when I was going to college my chemistry professor said, on her first day in class, "you are allowed to cheat in my class." "You can bring the books to my chemistry exams." "You can bring a calculator or a computer." "You can write the answers on your arm." But then she continued. "All my tests are timed. If you have to look up the right equation to use you will not pass." She was right. If you didn't do the homework and study your ass off and really learn the topic you weren't gonna pass. Asking ChatGPT for the answer is slower than knowing the answer already. Education needs a complete overhaul. They should start by studying my chemistry teacher. She knew how to teach.

@LuizaJarovsky "Rarely is the question asked: Is our children learning?"