here is o1, a series of our most capable and aligned models yet:
o1 is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it. openai.com/index/learning…
but also, it is the beginning of a new paradigm: AI that can do general-purpose complex reasoning.
o1-preview and o1-mini are available today (ramping over some number of hours) in ChatGPT for plus and team users and our API for tier 5 users.
May 13 • 11 tweets • 3 min read
live-tweeting our live stream in 1 minute!
desktop app and new UI
Mar 14, 2023 • 6 tweets • 1 min read
here is GPT-4, our most capable and aligned model yet. it is available today in our API (with a waitlist) and in ChatGPT+.
it is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it.
it is more creative than previous models, it hallucinates significantly less, and it is less biased. it can pass a bar exam and score a 5 on several AP exams. there is a version with a 32k token context.
Mar 12, 2023 • 11 tweets • 2 min read
ugh fine here is my SVB thread:
TL;DR: at this point, to be certain of avoiding catastrophe, the FDIC needs to temporarily guarantee all deposits. other solutions might work, but this is the best one.
first, this really is just a liquidity issue. depositors at SVB are going to get all or most of their money back, and will have a significant fraction of it this week.
Mar 10, 2023 • 4 tweets • 1 min read
dropping standardized tests while maintaining legacy admissions policies and claiming it's about advancing equality of opportunity is not a serious position.
i'd love to see data, but as imperfect as tests are, i bet they do more for equality of opportunity than most of the rest of the application process, and it's very possible to look at scores in context.
this is the echochamber gone very wrong.
Feb 19, 2023 • 7 tweets • 1 min read
the adaptation to a world deeply integrated with AI tools is probably going to happen pretty quickly; the benefits (and fun!) have too much upside.
these tools will help us be more productive (can't wait to spend less time doing email!), healthier (AI medical advisors for people who can’t afford care), smarter (students using ChatGPT to learn), and more entertained (AI memes lolol).
Dec 26, 2022 • 4 tweets • 1 min read
there will be scary moments as we move towards AGI-level systems, and significant disruptions, but the upsides can be so amazing that it’s well worth overcoming the great challenges to get there.
and we’ll see great benefits all along the way—they will make ChatGPT look like a boring toy.
Dec 3, 2022 • 6 tweets • 1 min read
interesting to me how many of the ChatGPT takes are either "this is AGI" (obviously not close, lol) or "this approach can't really go that much further".
trust the exponential. flat looking backwards, vertical looking forwards.
the field has a long way to go, and big ideas yet to discover. we will stumble along the way, and learn a lot from contact with reality. it will sometimes be messy; we will sometimes make really bad decisions. we will sometimes have moments of transcendent progress and value.
Nov 30, 2022 • 5 tweets • 1 min read
today we launched ChatGPT. try talking with it here:
chat.openai.com
language interfaces are going to be a big deal, i think. talk to the computer (voice or text) and get what you want, for increasingly complex definitions of "want"!
this is an early demo of what's possible (still a lot of limitations--it's very much a research release).
Apr 21, 2022 • 8 tweets • 1 min read
At the likely risk of really embarrassing myself:
Here are 2 from the latest version of GPT-3 trying to write startup advice tweets as me, 2 from @ryancohen25 trying to do the same thing, and 2 from me.
Can you tell which is which?
“Try something and iterate” is a better strategy than anything I’ve ever seen in a consultant’s slide deck.
Apr 10, 2022 • 7 tweets • 1 min read
If you are fortunate enough to work in the tech industry, you may well not have to work very hard to do well in your career.
Not true for everyone, and not true for lots of other industries, but plenty of evidence all over Twitter for this.
(Hopefully this will be the case for many more people over time as AI advances. We should aspire to a world where people who don't want to to work hard don't have to, and trust that society will find other ways for people to find meaning.)
Mar 20, 2022 • 11 tweets • 2 min read
I think US college education is nearer to collapsing than it appears.
Most of all, it's clearly a bad deal for many students, or we wouldn't have the student debt crisis.
Cancelling student debt is good if it's tied to fixing the problem going forward, which means not offering it, or having the colleges be the guarantor, or ISAs, or something.
Feb 4, 2022 • 9 tweets • 2 min read
Techno-optimism is the only good solution to our current problems.
Unfortunately, somehow expressing optimism about the future has become a radical act.
Take climate change. Some people seem to prefer shaming others into doing as little as possible and not having kids, instead of pushing for technological breakthroughs that can deliver limitless, cheap, clean energy.
Jan 22, 2022 • 5 tweets • 1 min read
Soon, AI tools will do what only very talented humans can do today. (I expect this to go mostly in the counter-intuitive order--creative fields first, cognitive labor next, and physical labor last.)
Great for society; not always great for individual jobs.
Using artists as an example, when anyone can create amazing art, there will be incredible upside for humanity, but downside for most individual artists.
(On the other hand, totally new kinds of art will be possible, and the skill that will matter will be imagination.)
Dec 1, 2021 • 4 tweets • 1 min read
Prediction: average venture returns for investments made in the 2020s are going to be much worse than those from the 2010s.
There's a flood of capital, and I have never heard so many price-setting VCs openly say they're willing to accept much lower return targets.
(Which I guess is sort of what is supposed to happen as the investing marketplace equalizes--the outperformance was just too dramatic.)
Nov 9, 2021 • 8 tweets • 2 min read
There are two ways to reduce economic inequality—you can work to lift up the floor, or work to lower the ceiling.
(Notice the extreme streak of authoritarianism that often accompanies the latter.)
Said another way, you can either believe
1) the pie can grow a lot (and so, e.g. we should encourage trillionaires and the creation of far more wealth), or 2) it can’t (and so, e.g., we should abolish billionaires and divide up everything we have now equally).
Nov 8, 2021 • 4 tweets • 1 min read
One of the most important career decisions you make is prioritizing status or substance.
Substance compounds and status decays.
Prioritizing status delivers quicker rewards but hits an early ceiling. Prioritizing substance often has a painfully slow start but no ceiling.
(The hard part is often taking an honest look at which is which.)
Oct 23, 2021 • 4 tweets • 1 min read
No matter what you think about Worldcoin, I think some of the ideas in here are likely important to the future of privacy: worldcoin.org/privacy-by-des…
I think Worldcoin is more privacy-preserving than centralized services we use today. All Worldcoin, or anyone, could ever tell is if someone has already signed up for the service.
The hash is cryptographically decoupled from the wallet and all future transactions.
Oct 23, 2021 • 4 tweets • 1 min read
Imagining an AI future (with some amount of redistribution) is an interesting lens for the Great Resignation.
Less people will have to work in the traditional sense and people will be less willing to do jobs they don't like.
People won’t have to work to survive, and we will have to pay more for less desirable jobs, or automate them, which seems great all around.
Oct 22, 2021 • 4 tweets • 1 min read
After 9/11 (3k US deaths), we got so much societal change and spent trillions of dollars.
After COVID-19 (700k+ US deaths), we seem to be getting very little change, and spending very little money.
I felt confident we'd get real change–competent planning for future pandemics, severe regulations on gain of function research, FDA reform, rapid-response vaccines for new viruses.
We are struggling to commit even a few tens of billions.
What's gone wrong?
Oct 3, 2021 • 4 tweets • 1 min read
Spoke to some really smart college kids recently.
Almost universally, they had a belief that climate change was an unstoppable disaster and the most pressing issue of our time.
The surprising thing was not one of them believed they could or should do anything about it; there seemed to be a total lack of belief in the ability to change the trajectory of the world.
