frontalpha.eth 🛡️
8.6K posts
frontalpha.eth 🛡️
@frontalpha_eth
Ethereum diehard and tech enthusiast
Katılım Nisan 2012
1.9K Takip Edilen2.3K Takipçiler
Yeah but of course these are correlated. use of margarine is a well known divorcable offense. This graph is not just correlational— it is causal.
Jill Filipovic@JillFilipovic
Can switching to butter save your marriage? As people in Maine have eaten more margarine, they also divorce less often. @lymanstoneky may be worth considering!
English
@bluewmist For me it was fitted sheets that go over a mattress
English
@benjaminsimon97 @dwarkesh_sp I agree. My thought was if Jensen doesn't think AI chips are equivalent to enriched uranium, that is also paradoxically bearish AI.
Is AI important or not? If it's not, sell it to China, but also, Nvidia stock should be valued much lower...
English
You don’t need to be a LessWrong rationalist (I’m certainly not) to agree with @dwarkesh_sp re AI/chip export controls and China.
Jensen’s position was frankly incoherent. “NVIDIA’s chips are far and away superior, the frontier requires NVIDIA chips” but also “China has plenty of chips and great researchers and can do it anyway, so why not at least have them build on American hardware.”
And: “the enriched uranium <> AI chips analogy makes zero sense” but also “America needs to beat China on AI.”
(Why is it important for America to win, if not because losing to China in frontier AI could have devastating economic and Natsec consequences?)
Incoherent and inconsistent.
magnus@magnushambleton
It’s instantly clear within the first 3min of the Dwarkesh Jensen episode that unlike every single other person that is at the center of the singularity, Jensen did not spend his early twenties debating things on LessWrong
English
@krishnanrohit I guess if we were to believe Jensen(why wouldn't we?) It's paradoxically bearish AI if he thinks it's okay to sell chips to China.
It AI isn't as exciting as enriched uranium, then this AI stuff is a bigger bubble than many people are ready for.
English
This reads confusing because it's at least 3 different arguments:
1) do export controls slow Chinese frontier model capability
2) do hardware ecosystems create long term lock-in
3) does China making a local stack weaken US global power as it diffuses
The claim that denying China Nvidia weakens US power is something like this:
- China is forced to build its own stack
- that stack gets good enough
- models and tooling optimise around it and developers and markets accrete around it
- US global platform leverage declines
Dwarkesh Patel@dwarkesh_sp
Distilled recap of the back-and-forth with Jensen on export controls: Dwarkesh: Wouldn’t selling Nvidia chips to China enable them to train models like Claude Mythos with cyber offensive capabilities that would be threats to American companies and national security? Jensen: First of all, Mythos was trained on fairly mundane capacity and a fairly mundane amount of it by an extraordinary company. The amount of capacity and the type of compute it was trained on is abundantly available in China. Dwarkesh: With that, could they eventually train a model like Mythos? Yes. But the question is, because we have more FLOPs, American labs are able to get to this level of capabilities first. Furthermore, even if they trained a model like this, the ability to deploy it at scale matters. If you had a cyber hacker, it's much more dangerous if they have a million of them versus a thousand of them. Jensen: Your premise is just wrong. The fact of the matter is their AI development is going just fine. The best AI researchers in the world, because they are limited in compute, also come up with extremely smart algorithms. DeepSeek is not an inconsequential advance. The day that DeepSeek comes out on Huawei first, that is a horrible outcome for our nation. Dwarkesh: Currently, you can have a model like DeepSeek that can run on any accelerator if it's open source. Why would that stop being the case in the future? Jensen: Suppose it optimizes for Huawei. Suppose it optimizes for their architecture. It would put others at a disadvantage. As AI diffuses out into the rest of the world, their standards and their tech stack will become superior to ours because their models are open. Dwarkesh: Tesla sold extremely good electric vehicles to China for a long time. iPhones are sold in China. They didn't cause some lock-in. China will still make their version of EVs, and they're dominating, or smartphones, they're dominating. Jensen: We are not a car. The fact that I can buy this car brand one day and use another car brand another day is easy. Computing is not like that. There's a reason why x86 still exists. There's a reason why Arm is so sticky. These ecosystems are hard to replace. Dwarkesh: It's just hard to imagine that there's a long-term lock-in to the Chinese ecosystem, even if they have this slightly better open-source model for a while. American labs port across accelerators constantly. Anthropic's models are run on GPUs, they're run on Trainium, they're run on TPUs. There are so many things you can do, from distilling to a model that's well fit for your chips. Jensen: China is the largest contributor to open source software in the world. China's the largest contributor to open models in the world. Today it's built on the American tech stack, Nvidia’s. Fact. All five layers of the tech stack for AI are important. The United States ought to go win all five of them. in a few years time, I'm making you the prediction that when we want American technology to be diffused around the world—out to India, out to the Middle East, out to Africa, out to Southeast Asia—on that day, I will tell you exactly about today's conversation, about how your policy ... caused the United States to concede the second largest market in the world for no good reason at all.
English
@lex_node Who would win?
1 million Bitcoiners screaming and foaming at the mouth: "THERE WILL ONLY EVER BE 21 MILLION"
or
sed -i 's/nSubsidy >>= halvings;//' validation.cpp && git push
English
I will support this if I never have to hear 'EtH dId the DaO haRdFoRk' again
newsflash: every single parameter of every single blockchain network is changeable, including 21M supply cap, all of it
ultimately you are trusting in other people (but a lot of them) & their ethos
Bitcoin Archive@BitcoinArchive
Cypherpunk Jameson Lopp and other Bitcoin developers propose BIP-361 to freeze quantum vulnerable wallets. This could lock dormant BTC like Satoshi Nakamoto’s 1.1M coins, now worth $74B, before quantum computers can steal them.
English
Maps the human nervous system's undocumented responses to specific frequency combinations of light and sound, discovering precise stimuli that trigger panic, compliance, disorientation, or paralysis in targeted individuals based on their unique neurological profile, delivered through any screen or speaker they're near.
English
Clone every IRGC commander's voice and issue false stand-downs. Fabricate perfectly contextualized evidence of betrayal, corruption, or back-channel negotiations with the US for every senior IRGC commander. Churn through exabytes of archived satellite imagery, cell tower logs, financial transactions, and intercepted comms to predict exactly where every IRGC commander will be based on a decade of pattern-of-life data no human analyst could ever cross-reference. Sensor algorithms that turn every camera into a microphone by reading sub-pixel vibrations on glass, every WiFi router into a radar that tracks bodies and heartbeats through walls, and fuses all of it in real-time across thousands of targets simultaneously.
English
Ok let's take your position that Iran fully reconstitutes in a year. The counterfactual is no strikes and Iran has operational nukes today! You're telling me you'd prefer the world where Iran has nukes right now over the world where, in your most optimistic scenario, they're starting from scratch and maybe get back to where they were in 12 months under sanctions, surveillance, and credible threat of re-strike? Does that make sense to you?
English
@frontalpha_eth @avidseries Their industrial base was never some robust high tech machine.
Just like the “destroyed nukes”. It will be back up and running within 1 year
English
Cue the repliers who think that decimating another nation militarily, even though it hasn't led to achieving the only two strategic goals that matter — regime change and the end of Iran's nuke program — means you've "won".
If the current situation is the ultimate outcome, the Iranian regime will, of course, have won, because it (and its high-grade uranium) will have survived. And, most importantly, it will have learned exactly how to fight the US in the future.
But my guess is that the current situation won't be the ultimate outcome because there's no way that Trump can sell it as a win except to the most retarded of his supporters.
What happens next is anyone's guess, but a continued battering of Iranian military, security and military-related manufacturing infrastructure could still produce an actual victory if the Iranian people become willing to rise up and shed the amount of blood that it would take to overthrow an increasingly weakened state.
Noah Smith 🐇🇺🇸🇺🇦🇹🇼@Noahpinion
America is the first country in over two centuries to lose a war to Iran, thanks to Trump
English
I did this with my baby. It felt very...like it was natural and right. It just seemed very effective and getting my daughter to eat. She could see I was eating it so she seemed more interested. Pre-chewing allowed her to digest food she normally couldn't. It's weird, but it worked for me.
English
@Romy_Holland I guess you could pre-chew meat like the Inuit and serve it to your baby. Kinda gross but that’s most likely a snapshot of what parents were doing 400 years ago.
English
i’m not a clean person and i am still horrified by the daily ritual of shoveling purées into my sons mouth with the full knowledge he’s going to spit 97% of it onto his chin, where it will then drip onto his highchair tray so he can mash it with his fat little hands for 5 minutes before i call it and hose him off in the sink. all this to get 6 calories of almond butter into him.
like, i just don’t buy that this is what ppl were up to 400 years ago.
English
@Romy_Holland I let my baby suck on tough pieces of steak she can't swallow because clean up was easier. She can suck out the nutrients, use her hands, feels good on the teeth, tastes good, poops aren't as smelly.
She was 6ish months in this video.
English
@83dollaroring It's easier for a large Army to walk than it is to swim...
People don't know this, but China's sphere of influence overlaps Russia's much more than the US's. Historically, China is a land power.
English
how r we feelin bout the “taiwan is a red herring and china’s actually gonna move on russia” theory
English
@sergeantsup My comment sounded funnier in my head. I don't actually know if this is true.
English
@frontalpha_eth I suspect this is not a common stance but feel free to show me evidence otherwise
English
most men could have much better photos on their dating profiles. and this would probably help them
but most "improve your dating profile" services have a vibe I disagree with. like recommending you become more hot in a really generic way
I guess it's similar to hotness advice for women
you can optimize for broadly appealing to a lot of people — getting your foot in the door many times
but finding an amazing romantic connection is about finding someone compatible with your weirdness. it's inherently an authentic and vulnerable thing
(if you hate wearing suits, do you really want to wear one in your dating profile? it'll attract women who want you to keep wearing it!)
there are ways to present yourself flatteringly *without* making yourself generic and inauthentic. you should use pictures with good lighting in your profile. pictures taken far enough from your face, so you're not distorted
you will get fewer dates than if you also optimize for hotness in inauthentic ways. but I think the dates you do get are more likely to be a good fit
English
@s_pect_re @JeffLadish Well, a powerful enough AI could push software updates to an anti ballistic missle network and make MAD obsolete. You would instantly be able to use nuclear weapons and be confident your enemy couldn't retaliate.
English
@frontalpha_eth @JeffLadish how and what do they dominate? this doesn't make any sense.
English
I hate to say it but an international agreement between the US and China to ban superintelligence is inevitable.
Leaders in these countries are just going to follow their incentives, and none of them are willing to give up control to an artificial superintelligence.
English
@martyrdison Women who are attracted to a full hairline are pedophiles.
English
i find it to be disgusting when men have thinning hair
if i see finanstride in their house, i turn away
bad physiognomy
should not be breeded with, excluded from the gene pool. especially if you can’t afford the flight to turkey
only thick hair creatures should procreate
English
@pegobry_en The only reason the Rafale exists as an operational weapons platform is because France is surrounded by friendly neighbors that bought the F-35's.
DEI ass weapons platform.
English
It's not nice to crow about the collapse of our British cousins, but you have to understand I have been chudding out literally my entire life (not my adult life, my entire life) about France having only one aircraft carrier. And of course in theory it should have many more. But within the constraints of Retardo-Europe, it turns out to have been a really wise choice by my elders. It turns out one really good aircraft carrier, with a nuclear power plant and catobar and a good naval group to escort it, beats the everloving shit out of two aircraft carriers you can't use.
Pascal-Emmanuel Gobry@pegobry_en
The UK building aircraft carriers that don’t work was fine by the US when it involved them buying F-35s they didn’t need instead of Rafale
English
@bluberino123 @alz_zyd_ I would argue if you have the compute to one-shot a better designed rocket, people are going to push the limit until you need to do 10 tests with a novel 0-1 rocket design again.
The hedonistic treadmill continues.
English
even if chollet is right and intelligence has a hard ceiling, human engineering is bottlenecked by bandwidth, not just iq. you don't need to break physics to launch a better rocket, you just need to stop paying for human mistakes. infinite intelligence means every single micro component is co-designed simultaneously by agents sharing data instantly. you get radically lighter materials, better fuel efficiency, and perfect tolerances on the first try. you aren't magically creating more fuel, you are just building a ship that doesn't need ten physical prototypes to avoid blowing up.
English
This is basically my view: infinite free intelligence will not hyperaccelerate science (though it may well put all the humans out of a job!) because science in 2026 is not actually intelligence-bottlenecked. We have basically enough intelligence already
François Chollet@fchollet
One of the biggest misconceptions people have about intelligence is seeing it as some kind of unbounded scalar stat, like height. "Future AI will have 10,000 IQ", that sort of thing. Intelligence is a conversion ratio, with an optimality bound. Increasing intelligence is not so much like "making the tower taller", it's more like "making the ball rounder". At some point it's already pretty damn spherical and any improvement is marginal. Now of course smart humans aren't quite at the optimal bound yet on an individual level, and machines will have many advantages besides intelligence -- mostly the removal of biological bottlenecks: greater processing speed, unlimited working memory, unlimited memory with perfect recall... but these are mostly things humans can also access through externalized cognitive tools.
English