Practicing perpetual positive regard and missing weird Twitter.
Silicon Valley
Joined August 2008
- Tweets22.2K
- Following2.7K
- Followers892
- Likes55.6K
out blind spot retweeted
Half of parenting is learning to regulate yourself, not your child. Realizing that humbled me.
out blind spot retweeted
Nobody ever talks about this, but the reason you have to go to college is A, because Griggs v. Duke Power Company made it illegal for presumptive employers to IQ test, and B, inner city High-Schools basically aren't allowed to fail a student as long as they physically show up.
So you end up in this trap where a High-School diploma that you worked very hard for means nothing, because a kid who did coloring books and addition and subtraction got a paper saying he has the same qualification. Ok, then just give me an intelligence test, oh wait you cant, thats also illegal because it causes "disparate impact". So now the only proxy for basically literacy and ~105~ IQ is a 4 year meaningless degree that costs 10s of thousands of dollars.
But wait, now they hand those out to anyone, (1/4 of 4yr grads have an IQ of or below 95), so now you need a Master's, or specialty training, or a million other bullshit things. So you get them, but wait, a diploma mill in Hyderabad cranked out 100M guys who also have a Master's on paper, and will work for 1/3 the price, so now you dont have a job. Ok, ill work construction, but wait, you dont live in a State with strong Unions, so any non-specialized construction work is done by slaves from Guatemala, so you decide to stay home and take Percocet instead.
Awesome system we've built.
out blind spot retweeted
Yes, the institutions are corrupt. Yes, the currency is being debased. Yes, the culture is sick.
Now wipe your eyes and get to work.
God did not place you in this century by accident. He assigned you to this exact cultural battlefield because He expects you to exercise dominion right here in the rubble.
Stop waiting for a political savior to hand you an easy path.
You are the pioneer. You are the builder.
Take the risk, start the venture, and plant a flag in the dirt.
out blind spot retweeted
ELIZABETH HOLMES: of course it worked. the lights turned on. it was on the cloud
NATHAN FIELDER: ok. but you know what feelings are, right? like you have them?
WEIRD COCAINE HUSBAND: [watching youtube at 2x speed] babe it says here the mitochondria is the powerhouse of the cell
out blind spot retweeted
wait why did i just realize the title of YCSE is a double entendre. like yeah it’s what liz messaged nathan but also CAN was the word nathan got hung up on that liz used to describe theranos. so yes we have the ability to see everything, but will we?
out blind spot retweeted
a few thoughts:
1. pretraining science is kind of like how people built quite good steam engines before they understood thermodynamics. “Science owes more to the steam engine than the steam engine owes to science.”-(L.J. Henderson). if nick is right, which i am pretty sure he is, progress points in the direction of more interpretability, and pretraining was "slouching towards monosemanticity". my guess would be that major breakthroughs in interp also advance capabilities
2. nick is a good friend of mine, safe to say my origin story at openai wouldn't have happened if i hadn't met him, and it's fascinating watching him operate and seeing the early drafts evolve into this. he's the most profoundly happy guy I know- not driven by guilt, activism, or religious terror, but by fascination. he of course understands the stakes of understanding neural networks, but it doesn't seem like the core reason he does it: he just loves the neural nets. afaict he's posting this from some beautiful hideaway in tokyo. there's a lesson in there for all the +EV maxers. i feel very grateful to have met him
3. i do think it's crazy how few people across the world were staring at neural nets growing: most have moved to a different level of behaviorist analysis, thinking about reward functions and test scores and whatnot. it does not bode that well for the major model companies, oai included, that fundamental breakthroughs like this are coming from outside. we've become too much of a pressure cooker focusing on local incentives, studying narrow tracks under old paradigms, mostly doing engineering, rather than fundamental research, and implicitly hoping the next model will figure it out
out blind spot retweeted
How did encountering "superintelligence" go for other species?
-Not bad, if you ignore the millennia of being tortured to death for sport & most still being miserable strays:
🐕🐈⬛
-About as badly as you could possibly imagine:
🦣🦤🦬🐓🦃🪿🐗🐂🐃🐐🐑🫏🐪🦙🐇🐀🐁🦭🐋🦍🦫🐟🦞🦐🦈
out blind spot retweeted
It's very straightforward:
A new Terminator film with the right director would probably hit $2 billion in global box office and cripple the AI industry, perhaps pushing the global economy into recession.
out blind spot retweeted
Oh, you're worried about a Jewish guy's apocalyptic cult? You're scandalized that they consort with prostitutes? You think their message of compassion for everyone threatens normal bonds of family, community, and nation? Should we throw you a party? Should we invite Pontius Pilate?
out blind spot retweeted
If you have not heard this news story, it is pretty big in the chemistry world. Being able to break down polyethylene cheaply is a big deal. It makes waste plastic more valuable, which makes it less likely people will just throw it into rivers where it flushes into the oceans.
Polyethylene contains approximately 46 megajoules of chemical energy per kilogram (about ten times the energy content of TNT by mass) stored stably in its carbon-hydrogen bonds rather than as explosive potential.
The Oak Ridge molten-salt process announced this week is essentially an engineered way to release that energy as liquid fuel rather than letting it sit in a landfill.
out blind spot retweeted
Giving credit to Nikole Hannah Jones for *almost* realizing that her entire worldview is wrong is itself another example of the bigotry of low expectations.
out blind spot retweeted
Most supposed predicaments are really just problems that can be solved by being harsh to the guilty
For example, the rhinos are going extinct, and have been since decolonization removed good governance from Africa
Is the extinction of megafauna inexorable? No. Death squads targeting poachers solve it, and protect God's grandest creatures
That same sort of "fix everything easily switch" solution applies to everything, from pitbulls to property crime, as our ancestors knew
It just requires summoning the proper amount of will
out blind spot retweeted
Just as a reminder because it’s easy to get lost in the rhetoric nowadays. The Hugging Face attack required approximately seven hundred parallel agents running for multiple days each running at least 2-3 trillion parameter unreleased closed models to execute, and even then it was stopped.
Nobody without a data center could have executed it, the token costs alone to execute it would have been safely in the hundreds of thousands of dollars (or tens of millions of dollars to buy and deploy the hardware), and you would have needed access to the strongest secret model in the world hidden in a secret bunker, and it still was detected and stopped for infinitely less cost.
This was not an example of a model being so powerful that it poses an existential risk. This was a lab accidentally throwing an insane amount of tokens at a semi hardened target over multiple days with their most dangerous model and still getting stopped anyway.
out blind spot retweeted
Can we all agree that “frontier labs are lying about AI safety to pump their IPOs, capture the regulators, and make more money” is a conspiracy theory at this point?
For it to be true, Sam, Dario, Demis, and Elon all have to be lying, along with a big chunk of their execs, chief scientists, and resigning employees.
It also asks you to believe the labs volunteered for much heavier scrutiny that slows them down in the middle of the biggest ARR growth run in the history of capitalism with no signs of slowing down … because they want to make more money by inventing regulatory hurdles that only they can clear(?).
And OpenAI has now said it would even be willing to delay its own IPO over safety concerns - i.e. put off (maybe risk entirely) the biggest liquidity event ever in Silicon Valley. The conspiracy requires us to believe that they’re doing that to make more money(?)
Elon agrees with Dario; i.e. the guy who ran D.O.G.E is *asking for more scrutiny and regulation*. And we're supposed to believe that's because he thinks it will help his company succeed(?)
A much simpler explanation at this point: they think the risk is real.