Google has a number of passion projects typically run by people who established credibility/value to Google a long time ago. Not surprisingly, some of those are medical/biological in nature, because that's an area that tech people like to contribute in after they've reached tenure.
I was a computational biologist and specifically went to work at Google to get access to their world-information-organizing technology to apply it to medical/biological problems. I was convinced at the time (mid-to-late 2000s) that AI was going to transform medicine, especially drug discovery, and that huge amounts of (organized) data was going to be key to achieving this goal. While there, we worked on protein folding and design and drug discovery, as part of a team that eventually was called Google Accelerated Sciences. It was mostly made up of people who had some level of scientific background, then had made $$$ for Google, and made good friends with the leadership and could use some of the research budget.
Of course, the protein folding and design work ended up being replaced by DeepMind's work on protein structure prediction, which led to protein design and drug discovery, mainly in DM spinoffs.
Many people at Google who work on CS stuff would absolutely love to see Google's resources applied to curing diseases. I know that Jeff Dean has been angel investor in this space.
IRB type completion comes as a result of a chain of events which starts from the incredible work done by Kevin Newton (et al) to write a new canonical Ruby parser called Prism in C99 with no dependencies [1].
With Prism, you can then create tool suites like syntax_tree [2], which then leads Prettier formatters [3], a new Ruby LSP [4], which unlocks a new Ruby LSP VS Code extension [5], not to mention a laundry list of other gems like Rubocop and of course Ruby itself that will benefit from a faster and more maintainable Ruby parser.
It's a beautiful illustration of the power of questioning conventions, going back to first principles to uncover better solutions to previously solved problems, whose new solutions create new capabilities which unlocks the ability to solve new problems.
syntax_tree was actually created before Prism. Ruby LSP also adopted syntax_tree first, and then switched to Prism (then called YARP) when it was mature enough.
But indeed, the type completor would've been much harder to build and maintain if without Prism.
If u have a ton money, it's cheaper to just live in US. Canada is very expensive for the plebs. The cost of living keeps rising and the politicians are very anti small businesses and pro oligarchy. They have a strangle hold on the country simliar to the Murdoch in Australia. Ironically, US is fairer in helping the small timers.
It was the saving grace, but I am not sure if that's the case now as the situation deteriorates. If I am already having the big tech insurance and I am gonna to fly to Mexico/India for some of the items anyway, why do I need it?
The politicians simply DO NOT care. They just let oligarchs keep eating the small palayers, and it will totally break the health system some day just like how they killed 3rd party internet providers. Telus is already making the play in the health sector.
As an H-1B holder, we are legally entitled (and required) to work for one specific employer and that's it. You cannot even think of starting a company, you cannot work a gig job, you can't work two "high-paying" tech jobs.
Everyone knows the "ton of money" doesn't come from working a salaried job, it's from creating something new / starting a company.
But only US green card holders or US citizens can even dream of considering that as an option they can pursue.
Ton of money is ill-defined, but it's quite possible to work a salaried job in tech in the US and achieve financial independence and an extremely high standard of living.
There was an internal joke / meme at Google that any announcement starting with "An update on X" == we are killing X, to the point that if someone was sending their resignation email the subject line of the email would be "An update on <name>"
My current favorite corp speak is “we can’t wait to share it with you”. Seems like it’s in literally every product announcement. At least this one is positive
Did it know that before the last LLM failure was posted on Twitter or Hackernews? Trawling tech media for LLM failures can be assumed to be part of the "human feedback".
Yes, the models are not constantly learning. They only update their knowledge when they are retrained, which is pretty infrequently (I think the base GPT models have not been retrained, but the chat laters on top might).
Crazy that despite their progress behind the scenes, they appear to have not touched this website since.
I probably spent a little too much time tweaking the CSS to get the mosquitoes to not overlap the text on various viewport sizes :)
reply