MTS
@MTSlive • 101,629 subscribers
Chronicling the singularity
Videos
AI is facing the same policy choice that defined the internet era. Adam Thierer, senior fellow at R Street Institute and a 30-year tech policy historian: "When I was growing up in the '70s, it was a big deal when we got a telephone that was a different color and had a longer cord. That was innovation back in the '70s." "Then the internet came along, we decided to have a firm firewall between that old era of regulated monopoly and the new world of digital technologies. We allowed them to be born free of government top-down control." "And we saw this flowering of entrepreneurialism and creativity and culture because of that policy choice." "Are we going to allow it to continue to be a technology of freedom? Or are we gonna bottle it all up and centralize?"
MTS160,997 Aufrufe • vor 5 Tagen
SITUATION EXPLAINED: How much are frontier labs actually spending on training data? .Sean Cai: "Frontier labs are spending about $10 to $15 billion per lab on data." "Really good long horizon tasks go up to $20,000 each. A complete browser-use version of SAP was rumored at $500,000." "Despite everybody thinking the market is super crowded, we still don't have enough good quality data vendors that actually understand how to deliver product plus services in a way researchers are looking for." "I have not seen a contract for genuinely good data gets turned down because of budgetary concerns yet."
MTS332,807 Aufrufe • vor 12 Tagen
.Adrian Dittmann is Zoomer-pilled. "If I were to be optimistic about one thing, I'm optimistic about Gen Z. I think they have the potential to become God's most locked in soldiers." "While they consume a lot of ridiculous stuff via the internet, it isn't their fault. In spite of this, they've proven to be an exceptionally resilient, informed, and capable generation." "It's not about the tools. It's about the people using the tools first and foremost." "You're better off training somebody on how to use their stick than to ban the stick... if they know what they're doing, any threat actor, enemy, or even a misaligned system will have no chance."
MTS122,280 Aufrufe • vor 6 Tagen
SITUATION EXPLAINED: What if Bryan Johnson's goal of not dying is actually rational? We asked Dr. Mike Israetel his thoughts: "Over the next five to 10 years, the vast majority of diseases and emergency things we couldn't treat, we will be able to. If you make it another 10 years, you could get to a point where we're reversing aging in the 2030s." "If someone told you, 'If you make it another 10 years, your probability of not dying is 100 times lower,' it sure would feel awkward to do stuff that's gonna increase your probability of death right now." "If Bryan Johnson was saying this in the 1940s, you'd be like, 'Dude, we all die, bro.' Nowadays, if you know there's gonna be hundreds and hundreds of years of your life after this, it can kinda change your perspective."
MTS139,001 Aufrufe • vor 7 Tagen
We asked the CEO of HuggingFace clem 🤗 what the risks of releasing powerful open source models are. He says restricting AI creates more risk than openness. "Six, seven years ago, at the time it was GPT-2, and there was already a lot of people saying that it was too dangerous to release in open source." "Mythos, when it was announced was crazy dangerous... In a few weeks or a few months, everyone is gonna be using Mythos, and not destroy the world as a result." "For cybersecurity, the biggest risk is that a few players have capabilities that other people don't have... If you make it more open, it's usually easier for defenders to react and make the whole system safer." "The idea of restricting a technology like AI based on risks is like saying, 'Some people can punch other people, so let's tie down everybody's hands.'" "Otherwise you slow down progress, you create massive gaps in terms of controls, in terms of capabilities, and you create actually additional risks."
MTS318,045 Aufrufe • vor 28 Tagen
SITUATION EXPLAINED: Why is every AI lab starting a deployment company? Matt Slotnick on what's driving the trend: "The thing we currently call FDE is gonna blossom into a lot more different jobs... all really about how do we bring applied intelligence into the flow of work and out of the data center and into the real world." "Every company will, in some ways, become a deployment company."
MTS122,679 Aufrufe • vor 12 Tagen
.Shaun Maguire says Elon Musk’s X acquisition was in service of civilizational-scale missions. "I truly believe, like in my soul, that he is trying to help humanity." "He has these goals. Like back up the biosphere by making us a multi-planetary species or ushering in electric vehicles." "These are incredibly important missions." "Politics became probably the biggest threat as a bottleneck to achieving these very positive missions." "I personally view the X acquisition as an attempt to help remove these bottlenecks. Through that lens, I think it's been incredibly successful."
MTS497,806 Aufrufe • vor 1 Monat
SITUATION EXPLAINED: is Universal Basic Income necessary? .Jordan Schachtel: "The history of technological progress has shown us one thing: quality of life, employment, and every metric that contributes to human flourishing improves." "Things might change, but preemptive UBI is insane. This is a multi-trillion dollar social welfare project, and we don't even know for certain there's a problem that it needs to solve." "The history is indicative of what happens. In the early 1900s, 50% of Americans worked in agriculture. 30, 40 years later, people were doing dramatically different things." "They were employed in jobs that required less labor, but took home more effective pay and had better lives." "We can't predict what happens (similar to the internet era, etc). But regardless, the answer is never to institute a giant multi-trillion dollar government program."
MTS124,480 Aufrufe • vor 13 Tagen
SITUATION ANALYSIS: AI data centers may be driving America’s nuclear comeback. What Christopher Koopman says is happening: • Nuclear companies are already building test reactors • Some are already going critical • Hyperscalers are demanding huge amounts of reliable power • Data centers need baseload electricity that can run around the clock • Chris says you probably don’t get the nuclear renaissance without the data-center build-out • The next unlock is small modular reactor rulemaking • After that, the question becomes how fast America can build
MTS153,301 Aufrufe • vor 19 Tagen
.Martin Shkreli: "The peptide thing is all BS." "Drugs are like seemingly simplistic. And the reality is, it's pretty freaking difficult." "Imagine the SpaceX landing, like the chopstick or the landing on its head... drug development makes that look easy." "Peptides in general are just this terrible asset class for pharma, because you need drugs that have long half-lives." "What this really is about isn't about pharmacology or science. This is about rebellion: I don't want to pay Pfizer. I don't wanna see my doctor. I don't wanna pay insurance... I'm gonna do this myself."
MTS409,404 Aufrufe • vor 1 Monat
What's the next big trade in AI infrastructure after memory? We asked Meltem Demirors and Kelly Greer: "Memory is ripping now. We think the next narrative is gonna be around networking." "Jensen is doing the exact same thing now that he did with Mellanox in 2019 that turned NVIDIA from a GPU company into a systems and networking company. He's doing it again with Marvell." "Hyperscalers are literally telling you the two biggest bottlenecks for AI spend right now are networking and memory." "The tea leaves are written literally everywhere... the market's just been on the memory track. We're trying to get ahead of that."
MTS34,843 Aufrufe • vor 7 Tagen
SITUATION EXPLAINED: Claude recovered Bitcoin worth about $397K after trillions of password guesses failed. X anon 🍜 bought 5 BTC for $200 in cash at a Starbucks in 2013, lost the password, and accidentally turned it into a decade-long forced hold. After brute forcing failed, he started feeding old notebooks and college laptop files into Claude. It found the missing connection between an old wallet file and a partially missing seed phrase. "I had two out of the three. Couldn't figure out the third." "It successfully worked." "Something I just never would've tried on my own."
MTS133,446 Aufrufe • vor 27 Tagen
We asked vitalik.eth why he frames Ethereum's broader mission as building "sanctuary technologies" rather than open source or decentralization. His take: "We're in a less peaceful and less safe world than we were 10 or 15 years ago. 10 or 15 years ago, you were worried about banks getting bailouts and dollars inflating. 15 years later, the risk that the dollar is gonna go crazy is way more credible ... and there's much worse things happening to people than just having your currency inflated." "There is a vision of safety we're competing with. Basically, let's trust the uncle in the sky, and the uncle in the sky is going to figure everything out for us in exchange for taking away all of our privacy and all of our agency." "We want to be safe and at the same time empowered. We want something that continues to keep us in control and at the center in a way where we have agency." sophia binji
MTS105,330 Aufrufe • vor 27 Tagen
.Andy Hall is bearish on UBI becoming a populist issue. " I don't expect UBI to take off anytime soon. "I think the politics of UBI are not favorable." "Americans are surprisingly skeptical of paying everyone in America money to do nothing. It's actually not a very winning political issue right now." "The best evidence we have on these randomized trials of UBI suggest they're not that successful at making people better off and kind of fix this sense of lack of purpose from joblessness." "So I actually don't think it's a winning political issue. I don't foresee it happening anytime soon. "
MTS77,821 Aufrufe • vor 22 Tagen
We asked Max Marchione how long he thinks lifespans will go to. "I think in the next 10 years we might be able to move the needle on lifespan to around maybe 130 reliably." "What we might see is it's very hard to break the biological cap of 140." "Of 60 billion people to live on Earth, we don't have a single documented case of someone living to 140, which is actually pretty crazy." "So we might see there's actually a cap and it's very hard to get beyond it, or we might see the opposite, which is, we're gonna create super intelligence, and super intelligence is gonna find ways to extend human lifespan to very large numbers." "Like, we're talking into the 100s, maybe even towards the 1,000s."
MTS115,015 Aufrufe • vor 1 Monat
What are the physics constraints holding back AI infrastructure? Meltem Demirors and Kelly Greer: "AI is basically just taking electrons from the ground and running them through GPUs, you get tokens out of your computer." "Power is the ultimate constraint, gigawatts of backlog to connect to the grid. How do we consume less power but spit out more tokens?" "When a computer consumes power, it generates heat. We're literally melting racks, Sam Altman tweeted about his GPUs melting at OpenAI." "The third constraint is information. It hasn't changed since Claude Shannon's seminal paper in 1941, we're still using the physics he uncovered then." "The next frontier is optics, computing done through light. But what's beyond the speed of light?"
MTS22,835 Aufrufe • vor 7 Tagen
.vitalik.eth on why a 1,000-year lifespan wouldn't get boring: "Even if we lived to 1,000 years old, we're not gonna get bored. We're gonna keep creating new worlds for ourselves." "Everything around us, the people, the world, changes so much over 10 years that it almost might as well be a death and a rebirth." "15 years ago, it was normal for close friends to not talk to each other for days. Doesn't that sound crazy now?" "In the first 20 years of your life you're a learner, a consumer. You're someone playing games that other people set up for you. One of the big transitions as you grow up is getting into more of a role of actually being the one that has to create and define and contribute to the games yourself." sophia binji
MTS83,668 Aufrufe • vor 27 Tagen