Comment by musicale
6 months ago
> The freedom to waste time. The freedom to waste resources. And the autonomy to decide how.
As the article notes, several companies (Apple, Google, etc.) could (currently) afford to fund such a lab, but there is no way their management and shareholders would approve.
There's a reason for this: research labs seem to benefit competitors as much as (or more than) the companies that fund them. This wasn't an issue for AT&T when it was a monopoly, but it is now. Personally I don't see it as a problem (since one home run innovation could pay for the entire lab) but company managers and shareholders do.
On the other hand, Apple does seem to have a de facto AI lab with a good deal of resource waste, so maybe that's good.
>> The freedom to waste time. The freedom to waste resources. And the autonomy to decide how.
> As the article notes, several companies (Apple, Google, etc.) could (currently) afford to fund such a lab, but there is no way their management and shareholders would approve.
Google did set up such a lab. The mission of Google Brain was literally to hire smart people and let them do work on whatever they want. ("Google Brain team members set their own research agenda, with the team as a whole maintaining a portfolio of projects across different time horizons and levels of risk." -- https://research.google.com/teams/brain/). Unsurprisingly, Google Brain is the place that originated the Transformer that powers the current AI craze (and many, many, many other AI innovations).
And they shut it down. In 2023.
The current tech giants spend a lot of money on "research," where research means optimizing parts of the product line to the 10^nth order of magnitude.
Arguably, Google Brain was one such lab. Albeit with more freedom than normal.
Which is fine, it's their money. But then they (and the broader public) shouldn't bemoan the lack of fundamental advances and a slowdown in the pace of discovery and change.
"And they shut it down. In 2023"
You mean they renamed it/merged it with another group that has similar freedom and focus on research
6 replies →
Interestingly enough, even non-monopoly large corporations once had labs where researchers had a good deal of freedom and where the projects were not required to be directly tied to business objectives. Hewlett-Packard, Digital Equipment Corporation, Sun Microsystems, Fujitsu, Sony, NEC, Toshiba, and Hitachi, just to name a few, had labs back in the 80s, 90s, and 2000s. As late as the early 2010s, a PhD graduate in computer science had options in industry to do research that wasn’t tied to short-term business priorities.
Unfortunately these opportunities have dried up as companies either got rid of their research labs or shifted the focus of their research labs to be more tied to immediate business needs. Many of my former classmates and colleagues who were industrial researchers are now software engineers, and not due to intentionally changing careers. Academia has become the last bastion of research with fewer commercialization pressures, but academia has its “publish or perish” and fundraising pressures, and now academia is under attack in America right now.
I once worked as a researcher in an industrial lab, but the focus shifted toward more immediate productization rather than exploration. I ended up changing careers; I now teach freshman- and sophomore-level CS courses at a community college. It’s a lot of work during the school year, but I have roughly four months of the year when I could do whatever I want. Looking forward to starting my summer research project once the semester ends in a few weeks!
> As the article notes, several companies (Apple, Google, etc.) could (currently) afford to fund such a lab, but there is no way their management and shareholders would approve.
When I was at Apple for several years, there were definitely at least two such groups.
Recently? Apple's Advanced Technology Group seems like it was fairly successful (QuickTime, HyperCard, ColorSync, AppleScript/Apple Events, Squeak, etc.) but Steve Jobs shut it down in 1997.
https://en.wikipedia.org/wiki/Apple_Advanced_Technology_Grou...
it morphed and part was in Oregon, part in Cupertino, certainly through the 2000s and 2010s. the name internally didn't seem changed but maybe it was slighty altered.
This is why I think we need publicly funded open source projects with paid leads. There are so many basic things we've failed to do for ourselves and our fellow human beings.
For example, the best non-AI TTS system is still Ivona TTS that originated at Blizzard in like 2007. The best open source solution is espeak and it's permanently stuck in 1980... Ivona was bought up by Amazon and now they don't even use the original software, but do charge money per word to use the voice via Amazon Polly. They could open source it, but they don't.
We don't even have something as basic as text to speech freely available, whether you are disabled or not. That is a problem. You have this amazing innovation that still holds to this day, squandered away for nothing.
Why can't we just have an institute that develops these things in the open, for all to use? We clearly all recognize the benefit as SysV tools are still used today! We could have so many amazing things but we don't. It's embarrassing
Google has DeepMind, Microsoft has Microsoft Research, Meta has FAIR.
It’s not trivial to foster such environments, but they do still exist in different forms.
They seem more applied, with the possible exception of MSR. None of them seems like the "check in a few years later" sort of place.
Facebook dumped $60B into an AI universe and HN made fun of them for it.
AI universe? Is that metaverse or something else?