ThomPete Paul Graham People who "work intensively with AI" could not predict GPT-4 capabilities. E.g. Yann LeCun have boldly claimed that even GPT-5000 won't be able to solve task which require innate understanding of physics in Jan 2022. The specific task he provided was solved by GPT-3.5 only a

Which quadrant do you live in? Statistically speaking my bet is the North Eastern one. Map shows the share of the global population living in each equatorial and meridional quadrant. Just a bit of geography trivia for you. Source: buff.ly/45dNIKa

Before your pen whispers a word to the paper, pause and ponder - 'Can #ChatGPT craft this tale for me?' #AI #DramaticEfficiency

“Yesterday is gone. Tomorrow has not yet come. We have only today. Let us begin.” ~ Mother Teresa

Geography trivia for you: The US records net positive migration from the entire world except Australia. Does that mean US Americans think Australia is the only more desirable destination on the planet? 😉🇺🇸🇦🇺

Learning from failure is surprisingly hard - in fact, it seems to happen rarely. So, have postmortems on what a team is doing right. This study where soldiers were either debriefed on failures, or failures & wins, shows including wins led to richer discussion and better outcomes

The pace of AI doesn't stop... Huge news this week from Adobe, Google, Meta, GPT-4 Vision, ElevenLabs, LLaVA, Anthropic, Geoffery Hinton, Harvard, OpenAI, Replit, UCL, TikTok, and the U.S. Space Force. Here's EVERYTHING you need to know (a thread):

Patrick Collison, Stripe CEO, on why quality matters: "If Stripe is a monstrously successful business but what we make isn't beautiful and Stripe doesn't embody a culture of incredibly exacting craftsmanship, I'll be much less happy. My intuition is that more of Stripe's

Ironically, out of 1,016 occupations analyzed for the level of exposure to AI (that is how much AI will overlap with the job, not that AI necessarily replaces that job), roofing is one of only a handful that will not be impacted by generative AI at all. arxiv.org/pdf/2303.10130…

“Beauty is truth, truth beauty,—that is all Ye know on earth, and all ye need to know.” Keats was right about our instincts. This paper shows we view more beautiful infographics as more trustworthy, and they get more “likes” & more scientific citations. osf.io/preprints/psya…

from FT

चाह गई चिंता मिटी मनुआ बेपरवाह। जिनको कुछ नहीं चाहिये, वे साहन के साह। ~ rahim

The 12 things that make a theory interesting, as classically listed by Murray Davis in a 1971 article, also can serve other purposes: like deciding what is a good topic for a popular article, or as a guide for finding disruptive ideas for the future. …ssnationalstudies.files.wordpress.com/2009/11/thatsi…

Saul Steinberg (1914-1999) - Untitled

Paul Graham Steve McGuire You can actually tell that across all colleges, Math majors are obviously better quantitatively but also better verbally than half of humanities majors.

When the first iron bridge was made (creatively named Iron Bridge) in 1781, no one knew how to build with iron, so the builders used traditional woodworking methods, like dovetail joints Thats the place we are using AI now, replicating old approaches until we figure out new ones

For of all sad words of tongue or pen, The saddest are these: ‘It might have been!’ ~ J.G. Whittier

How big is Africa, in reality.

I screenshotted this because it will seem so funny a couple years from now. Though to be fair it seems pretty funny already.

SpaceX “NASA has this phrase that they like: ‘Failure is not an option.’ But failure has to be an option in art and in exploration, because it’s a leap of faith. And no important endeavor that required innovation was done without risk. You have to be willing to take those risks. So,