OpenAI, the nonprofit research organization founded by Elon Musk, has bested a world champion e-sports team OG in Dota 2.
OG faced off in a best-of-three contest against the OpenAI Five bots, all trained using the same deep reinforcement learning techniques and controlled independently by different layers of the same system. Reinforcement learning is effectively a trial and error approach to self-improvement, wherein the AI is dropped into the game environment with zero understanding of how the game works and trained extensively using reward systems and other incentivizing mechanisms.
Evernote has been plagued by a series of managerial missteps and failed product launches. It was one of the first apps in the App Store and the first notes app with cloud sync but it lost its battle to the new entrants as the whole tech stack has become commoditized. Reading that article about the rise and fall of Evernote and as I learn more about the deeper vision I think calling Notion Evernote 2.0 is actually fair.
Given the current revenue multiple of Zoom (which is absolutely batshit crazy high and awesome), is it safe to conclude that the market is very desperate for profitable tech IPOs in general and is losing sanity over this one company because its so rare?
12:53 PM - 18 Apr 2019
A whole tweetstorm from Sarah Guo, Partner at Greylock, here.
1) System is so broken up that all innovation has stopped 2) Create a parallel system just to make any improvements 3) Even though the new system is in its infancy it still starts winning 4) Tweet “lol we already had [the system] that’s not innovation” https://t.co/RrGdnr46a9
Wow, IBM had a 20% equity stake in Intel in 1982 but sold out in 1986 for $625m. Ten years later that stake would have been worth $25bn. IBM also had the option to buy 30% of Microsoft in 1986 for <$300m. They didn't. Even by 1996 that would have been worth $33bn+.