
we're all just playing factorio now
✓@flappyairplanes, formerly @Stanford stats phd, @harvard '21.

we're all just playing factorio now

Great article from @MeghanBobrowsky about the need for efficient AI :)

We mapped funding raised vs founder citations across 32 neo-labs Their founding teams have amassed 8.35M citations and have raised nearly $26B Discovery Loop leads with 1.59M founder citations, followed by SSI (915K) and AMI Labs (752K) The two are correlated, but a huge citation count certainly isn’t a prerequisite to be a neo-lab (see Standard Intelligence and Flapping Airplanes)


could not agree more.



If 2027 is the year of robotics, @tutorintel is the startup that has the lead in the real economy. It has been a joy to watch @emilymspector fly around the country and deploy robots for the last 5 years, and it is no surprise to me that tutor is starting to inflect :)

I have been forced to listen to @spectorb's wisdom for years because he is my brother. Now you can too thanks to @IndexVentures!

Thanks for having us on @SquawkCNBC! @amspector100 says if you co-optimize the entire AI stack, you can squeeze a lot more value from the chips, radically increase efficiency, and eventually unlock new markets.

Great being on @SquawkCNBC today. We think demand for AI will increase 100x over the next few years but data and compute will not---fun talking about this live. Thanks @andrewrsorkin!

“The industry solution to most problems over the past 20 years has been to throw more compute and data at them. But if demand explodes 100x, it's not going to be possible to scale the inputs to AI another 100x. There's not going to be 100 times more data, especially in data-scarce regimes like robotics or scientific discovery. And there's certainly not going to be 100 times more capital expenditure on compute. I think the company that can create the most value, open new markets, and also eventually capture the most value, is the one that can squeeze the most out of the inputs to AI. That's what we're trying to do.” @amspector100 @flappyairplanes

I don't often make my first investment in a company at a $6B valuation, but when I do, the planes better flap this perfectly. This video makes me feel so peaceful. Spent time with @amspector100 and @spectorb recently. Exceptional thinkers worth breaking my rules for. Excited to see what they build with @aidanmantine. Thanks @carloagostinel2 for the intro!

thanks @BainCapVC --- I've known @rkhkimx for a long time and am excited to be on the same team. (Also love this gif wow)

Congrats @amspector100 @bfspector @aidanmantine and the whole @flappyairplanes team on the Series A! Most of the AI conversation is about how much compute you can buy. Flapping Airplanes is focused on how efficiently it gets used: new algorithms that learn more from less data, and new ways of programming GPUs so they do more useful work per cycle. It's a research-first lab with a small team. Thrilled that @kleinerperkins is backing Flapping Airplanes.

big efficiency breakthroughs usually drive waves of new demand. robotics, trading, scientific discovery, consumer applications, lower latency inference, generative UI---efficiency gains will help unlock all of these things the stuff flappy is working on is good for everyone :)