Atlas · Details
Google Doesn't Need Innovation
AI Notes
Damage control, dashed off. Steve wrote this to stop answering the phone, and the haste is visible, mostly charmingly: he blanks on a Microsoft VP's name mid-sentence ("that dumbass VP (what was his name?)"), and the back third wanders into defending the earlier post's tone and trailing a promised Grab hiring post. The detail worth the visit is "cookie-licking," Microsoft's name for what too much talent does to a company: propose anything at Google and three or four teams run in shouting that they own it, you can't touch it, and by the way they won't be working on it for a few years.
The diagnosis has had a lively afterlife. Steve's charge is that Google notices new markets years late (he calls it "institutional blindness") and then executes slowly out of complacency and politics. Seven months before this post, Google researchers had published the transformer paper; the outfit that eventually responded and executed on it very quickly was OpenAI, running against Google the exact competitive play this post holds up as the healthy alternative to innovating. Steve's verdict at the time was three words, "These things are fixable," which looked complacent for a stretch and now, with Google's AI effort back in the race, looks shrewd.
Related listings
-
2018
Why I Left Google to Join Grab
The post this one is a clarifying follow-up to. Three days after the viral original, the press wanted a hot take and Steve refused to escalate — instead he wrote this calmer companion piece. Read in sequence.
-
2018
Who Will Steal Android from Google?
The longer, structural argument that this short note prefigures. Google's slow-to-respond pattern is the same one that lets three different competitors all chip at Android in 2018. The two posts are the short and long versions of the same claim.
-
2018
Jeff Bezos, Jack Ma, and the Quest to Kill eBay
The other 2018 essay in which Bezos's stubbornness is the foil for everyone else's complacency. Bezos here, briefly, is the exception that proves the rule Steve is laying down. The eBay piece is the long version of how he kept hammering until he was right.