Tweet

See new Tweets

Conversation

Ethan Mollick
@emollick
The assumption that AI will soon run out of data to train on, or that it will become “corrupted” by training on AI-created content, is just an assumption. There are already cases where AI trains itself on AI data, and the limits of these techniques are not yet clear.
Quote Tweet
Arvind Narayanan
@random_walker
A popular view. I get it — learning w/o human knowledge seems implausible, even insulting. But let's wait and see. We already know this is false in some domains like game playing — synthetic data *alone* (self play) yields superhuman agents. In which other domains can it help? twitter.com/tdietterich/st…
17.5K
Views
11
Retweets
3
Quotes
60
Likes
8
Bookmarks
David Watson 🥑
Reply
christian
@curious_vii
if humans can improve our world models through thought experiments, then shouldn't AI models be able to get better with synthetic data? of course, this only works up to a point w/o real-world experimentation, but reasoning & reflection seem to get us pretty darn far!
1
1
173
christian
@curious_vii
also, I think that the value of having a high-fidelity private digital repo of your life (a la
@RewindAI
) will soon be so great that there WILL be lots more data to train on...just a matter of each person deciding what to keep secret vs. give up to foundation model developers
2
853
Show more replies

More Tweets

Kirill So
@kirso_
Product Management is easy! All we need to do: Is create wonderful products for users And launch it on time And make it sustainable for the business And make engineers understand the solution And then verify if they've built the right solution And convince 100 people to… Show more
3
242
Navah Wolfe 🚀🚀
@navahw
My 10yo child, upon disagreeing with a parenting decision: “I understand this as much as I understand Icelandic. But it’s different, because even if I don’t understand Icelandic, lots of other people do. NO ONE else understands this decision but you.”
16
647
18.2K
Tristan Homsi
@homsiT
"hell yeah or no" (originated by
@sivers
?) is the most important truism in startups. applies to hiring, product ideas, partnerships, software refactors, etc. you just get beaten down again & again if you settle for anything mid
20
1,220
Matthew Brown
@MCovBrown
I'm so hyped for my wife. Started a service business. Sold first customer 3 months ago. Now sold ~150 customers over the phone. Runs and operates the entire thing remotely. Just hired a VA. No previous business experience. Lesson: take action & you can learn so much so fast.
16
151
33.8K
WMP
@WPeaster
One of my bucket list to-dos? Vote in
@nounsdao
proper! Now I can thanks to
@WilsonCusack
delegating Noun 652 to me for this July 🙏 My platform? Hyper pro-spending, hyper pro-experimentation. Will be posting my vote reasonings on the TL here over the next few weeks ✌️
4
40
2,306
Minn Kim
@minney_cat
My family and I moved here to build our American Dreams 🇰🇷🇺🇸 Embracing immigration is to welcome the gift of global talent to transform the US If you want to navigate US immigration w more ease and speed,
@lisawehden
and I are working on this and would love to chat. DMs open
Quote Tweet
Sam Altman
@sama
one of the easiest policy wins i can imagine for the US is to reform high-skill immigration. the fact that many of the most talented people in the world want to be here is a hard-won gift; embracing them is the key to keeping it that way. hard to get this back if we lose it.
1
144
42.5K
Aaron Levie
@levie
The power of AI is it’s a lever on human time and expertise. The real potential is not replacing stuff we already do, but solving the problems we could never get around to because our time is scarce and expertise not evenly distributed.
18
265
41.8K
Dobroslav Radosavljevič
@dobroslav_dev
I thought that
@Vercel
would internally hate me for tweets like this because it got a lot of attention (this and a previous tweet about Next.js) But they actually DMed me (including the CEO) to help me solve the issue. I partially solved the compiling times by upgrading to… Show more
Quote Tweet
Dobroslav Radosavljevič
@dobroslav_dev
Today is the last day I'll use @nextjs with App Dir... The compilation is so slow, that I can take a coffee break between saving file changes. It's just not possible to develop something without going insane, sorry. 10s for compiling 1 line of code change is the worst DX. I… Show more
27
297
202.7K
kepano
@kepano
File over app File over app is a philosophy: if you want to create digital artifacts that last, they must be files you can control, in formats that are easy to retrieve and read. Use tools that give you this freedom. File over app is an appeal to tool makers: accept that all…Show more
45
1,825
329.2K
Jim Fan
@DrJimFan
It’s fascinating that somewhere in Princeton’s fruit fly brain scan lies Nature’s algorithm of attention, working memory, and even basic consciousness. Think of the connectome as a compiled executable binary. How much source code can we decompile from a full brain simulation? 🤔
read image description
ALT
Quote Tweet
Sven Dorkenwald
@sdorkenw
0:08
We are releasing a whole-brain connectome of the fruit fly, including ~130k annotated neurons and tens of millions of typed synapses! Explore the connectome: codex.flywire.ai Reconstruction paper: biorxiv.org/content/10.110 Annotation paper: biorxiv.org/content/10.110 1/6
22
938
227K
John Carmack
@ID_AA_Carmack
Has anyone ever made a PC VR spatial audio setup that used fixed speakers in a room so no headphones are necessary? Ideally apps would directly mix for the speakers, but you could also “reverse spatialize” head tracked headphone audio to world locked speakers.
41
488
114.5K