That was a crazy period when Google first fired him, then fired people criticising him, and then fired people criticising the people that just got fired.
But let's be honest, the guy was kind of unhinged. I would not have fired him, but neither would I have kept him in my team.
I agree, internally, looking at a fuller picture of his activity, he was off. Constantly bringing the subject up out of context, starting fights, etc. but they should have just warned him.
I'm sure he's just overjoyed to be tried without representation or evidence in the court of public opinion, a big step up from being punished with no process at all.
He wasn't unhinged, he somewhat clumsily posted evolutionary biology literature fragments on a channel where it would offend parts of the readers.
In response to a "let a thousand flowers bloom and speak your mind" request from Google management snakes. The problem is that some tech people take these requests seriously.
Google of course has identified itself as Trump sycophants and hypocrites by now. Maybe they should invite Jordan Peterson, Gad Saad and Elon Musk to give keynote speeches.
That is hard for outsiders to know. Google however seems to foster cancel mobs.
Python core people like Thomas Wouters and Gregory P. Smith, since fired from Google, canceled and libeled people like Tim Peters. All that matters is having a mob on your side and preventing the other side from responding.
What other things did he do to make you believe he was “kind of unhinged”? I know about the memo thing, but I think it was covered in an inflammatory way because the press loves to play up a controversy. What else besides the memo incident did he do that merits his firing?
Damore had been aggressively arguing about this stuff internally for some time in advance of his doc. The doc was consistent with a pattern of "makes a huge stink of things in every training and refuses to back down."
Okay, here's what I still don't understand. Were the things he was “making a huge stink about in every training” true or false? Reports at the time claimed that he was ”perpetuating harmful stereotypes,” but other reporting claimed that he was pointing out facts that inconveniently challenged the cultural orthodoxy and then wrote the doc to clarify his feedback, and so he became a target for internal activists. Are you aware of any specifics that would allow a reasonable person to lean one way or another as to the reality?
By all means, explain what else he did. I've seen this claim made repeatedly on hacker news, but conspicuously noone seems willing to o substantiate those claims.
One thing I didn't mention in the story is the potential consequences of top RU hackers being outed [...] such as interference and shakedowns from tax authorities and local police, extortion or even kidnapping for their considerable ill-gotten crypto wealth.
The TileRT approach swaps throughput for latency, which also means less overall efficiency
Given the export restrictions this could mean they need to prioritise how to best use their limited hardware. But they could also be moving to Huawei GPUs like deepseek did and simply not have stable hardware or software for a large scale deployment yet.
This is just speculation based on the MXFP4 support on Huawei GPUs that is lacking on some nvidia GPUs.
Because (1) Huawei collab and (2) vLLM etc dont implement half of the inference optimisations deepseek proposed in their paper.
reply