Twitxy Saved

nwyin

Verified account This account is verified. Learn more

@_nwyin

per aspera ad astra
nwyin.com/where nwyin.com Joined July 2018
267 Following 642 Followers 3.1K Posts

Latest posts 5

Pinned post

nwyin Verified account @_nwyin
pay more attention to what you want to see more of.
you are permitted to do what you believe is right.
risks are less risky than you think. take as many as you can.
the more you know, the easier it is to see the next steps.
if you’re ever confused about what to do, do the right thing.
you haven’t failed until you’ve given up.
expectation of progress towards a goal is key to motivation.
let people tell you no. don’t make the decision for them.
decisions autoresolve to the scary one.
strong intuition doesn’t mean intuition is right.
seek ground truth and poke reality. don’t settle for proxies.
send cold emails. assume that everyone is a friend.
be suspicious if you haven’t felt awkward today.
working on hard problems requires courage. acknowledge & embrace fear.
the smarter someone is the more they can afford to have terrible epistemics and still be successful.
if there’s something on your mind, write it down & get it out. maintain full attention on what you’re doing.
if you have a thought and you don’t like it, you can tell your brain that you don’t like it and drop it.
avoid getting stuck.
if you never feel a particular emotion, it’s probably the brain suppressing awareness of it or suppressing the emotion directly.
read. a lot.
write. even more.
the richest people lose the most money, the most connected people get rejected the most, the most successful people fail the most.
people think that others have same motivations as them.
people protect themselves by projecting their fears on others.
what’s the default decision?
run, don’t walk. literally. there is no speed limit.
you’ll learn 10x more by talking to the paper’s author for 30 minutes versus reading it for 30 minutes.
you can’t improve what you can’t measure. if you want to improve on ground-truth, measure ground-truth.
you think you want something until you meet someone who actually wants it.
we become the people whose opinion we care about.
we become the people we spend the most time (physically and mentally) with.
get to the point.
how do you feel – not think! – about your current top goal?
fewer, better arguments.
cherish friends who tell you things that make you angry.
easy to replace systems get replaced by difficult to replace systems.
do something courageous.
be suspicious if you’re not asking dumb questions.
feeling stupid now is better than feeling stupid in 10 years.
some people are not your friends.
saying things out loud shows inner conflict.
make sure you trust your own word.
writing makes thought visible.
luck favors the prepared.
everything is power-law distributed.
everything is power-law distributed.
if you’re not failing, you’re not operating at the edge. if you’re not operating at the edge, you’re not learning as much as you can.
ask for help. especially when you’re scared of asking for help.
grow a personal website.
if someone you respect gives you dumb advice, think about it instead of discarding it immediately.
let the world tell you no.
have a concrete, positive vision of the future.
seek people who don’t get angry at you for saying what you think, being yourself, and doing what you believe is right.
love.
you decide whether to live in the past or in the future with your every action.
spend more money.
know what you want.
be loyal to your friends.
ask the girl you like out.
be gentle. be kind.
don't die.
nwyin Verified account @_nwyin
coordination probably isn't the main bottleneck right now

we doubled population since 1970 and created the Internet and Slack and Linear yet...

what do we have to show for it?

LLMs who can do more than talk over each other like Redditors will be important, of course, but it's not really holding back current economic growth. if you could get all white collar workers to adopt it today i'm sure we'd see a blip of GDP growth that's at least 2x what we have now

but for enduring acceleration, there's still more progress to be made on model capabilities (longer horizon coherence, memory, continual learning)
a16z @a16z 11h
Google DeepMind's Alex Imas says AI agents still struggle to coordinate, and that's a key bottleneck to economic transformation:

"What you actually need for real economic transformation is not just competence... It's the ability of these agents to coordinate with one another."

"There's this idea of agentic swarms, right? But if you unpack it, these agents are still very bad at coordinating. They're not coordinating, they're mostly just doing parallel processing."

"If you have benchmarks specifically on coordination, you could actually show that 1 agent does better than 2 agents splitting the work, even if they can communicate. Whereas if you give the same thing to people, they actually have division of labor, they communicate, and they do way better than 1 person with double the resources."

"But if they get better at coordinating, you can basically break the Coasean ceiling on firm size, and you can get very, very different firms to emerge."

"Once coordination becomes better, the economic space of possibilities of what firms look like, what happens to wages, what happens to employment, that's when the real kind of visible changes are going to happen."

@alexolegimas
View quoted post on X