5 ms·
“Trust me bro” is not a reliable source, no matter I how important or famous someone is. Edit: believing one’s anecdotal story is different to accepting the or
by noncoml 4y ago
“Trust me bro” is not a reliable source, no matter I how important or famous someone is.
Edit: believing one’s anecdotal story is different to accepting the original assertion of the OP as true. Not saying they are lying, just take it with a pinch of salt.
- dekhn 4y agoFeel free to disbelieve me. But I think I got a pretty good read on the situation having talked to a wide range of the original players and a fair number of xooglers and googlers who frequent this site have probably interacted with the old leadership enough times to know what their real motivations were.
- ocdtrekkie 4y agoHow do you feel Google today squares with the Larry Page you met who said advertising funded search engines were inherently unethical?
- dekhn 4y agonot a fan of the business itself (IE, Sundarland) but I continue to appreciate the amazing technical work of many of its employees. I wouldn't put too much stock into what people wrote about what Larry said back in the early days; it's a lot more heat and noise than light.
- atlasunshrugged 4y agoHave you kept up with them? I know Deepmind is still a leader in the space and part of Alphabet now but I never hear much about Larry engaging in AI work (I hear much more about Sergey's airship or Schmidt's efforts warning the U.S. about a rising China).
- williamscales 4y agocursory googling (hah) shows me that dekhn appears to have been employed at google doing research work on protein folding. so i think it’s pretty plausible.
- dekhn 4y agoYes, that's correct (although I mostly built the infrastructure for exacycle rather than doing the actual research). At one point I personally had control over more CPU cycles (choosing which jobs to run) than any other person on the planet. I also worked on the Ads Database and its related data processing systems, plus the Making in Science team (we built and ran the booth at the Maker Faire), created the predecessor to Google Accelerated Science, and spent years working on ML software and hardware. My last work, unpublished (and probably gone forever) was using the idle time on TPUs to establish the using multiple large TPU and standard clusters to do multitask training on large multimodal corpus (the docjoins, labelled youtube) with the goal of creating a model that generated a image of a human head that could be interacted with verbally, and had enough context/state to pass any test humans give it (my prior is that you can make a non-sentient system that fools experts, not Blake, although I would be delighted to learn that merely making such a system generated a truly sentient life-form, it seems like that would be truly impossible to demonstrate convincingly).
- astrange 4y agoThis is a good website for actually believing someone, I think. (In the last year or two I posted an otherwise unreported anecdote involving a famous tech person and the only response was someone telling me I made it up.)
- karaterobot 4y agoThey're saying they were an eyewitness.
- pigtailgirl 4y ago-- isn't taking someone at face value one of the more valuable merits of this community? --
- summerlight 4y agoI don't know who dekhn is, but I can confirm that lots of the factual stuffs explained in the parent comments are true as an engineer working on heavily ML related stuffs.