Where are the main AI Labs?
This week an interesting report came out from The Brookings Institute about artificial intelligence. It was written by Mark Muro and Sifan Liu and was titled, “The geography of AI: Which cities will drive the artificial intelligence revolution?”[1] You can go download the report which is well put together or they even have a shorter executive summary you could grab instead.[2] My suggestion is to go read the full report. It is worth the time and investment. We are getting a lot of reading material coming out recently and I know the machine learning space is full of a never ending stream of new content. Make sure to go check out page 13 of the report. It has a map that shows, “AI employment concentration by U.S. metropolitan area.” You will very quickly realize that a group of companies are really driving all that hiring and they are covered in table 2 at the bottom of page 15 of that report. It's the usual suspects for geographical location, “These include eight large tech hubs—New York; Boston; Seattle; Los Angeles; Washington, D.C.; San Diego; Austin, Texas; and Raleigh, N.C.—and five smaller metro areas that have substantial AI activities relative to their size: Boulder, Colo.; Lincoln, Neb.; Santa Cruz, Calif.; Santa Maria-Santa Barbara, Calif.; and Santa Fe, N.M.”
All right, that epic pile of research on the geographical intensity of jobs in the artificial intelligence space aside, let's dig into the main question for the week, “Where are the main AI Labs?” One of the interesting things about this question is if groups like EluetherAI and Hugging Face should be included. For the sake of simplicity I’m going to acknowledge that I thought about that and move along to the bigger named players who are typically considered when you think about AI Labs.
The Google Brain team might be officially located out of Mountain View, California, but they have satellite groups all over the world.[3] Strangely enough they don’t have a group in Boulder or Denver which was a little surprising. Don’t forget that DeepMind was acquired by Google back in 2014 and has research centers in Canada, France, and the United States.[4] Over at OpenAI they are headquartered out of San Francisco, California.[5] After doing a little bit of research I begrudgingly need to include the location of FAIR (Facebook AI Research).[6] They are located out of Menlo Park, California, Seattle, Paris, and NYC.[7] For the most part you can see that the main artificial intelligence labs where people are working together have some pretty similar locations. That report mentioned up in the first paragraph from Brookings pretty much confirms that speculation with a reasonable dataset. The next wave of collaboration on some really interesting things is happening outside of that from some of the smaller open source projects and efforts. Those efforts have people from all over the place working on them and given that we are in the golden age of people training and learning about artificial intelligence that is only going to expand exponentially until the nature of the effort changes.
I read this article from The Conversation which was pretty good about how, “Google and Microsoft are creating a monopoly on coding in plain language.”[8] It made me wonder if a major advance in plain language coding will democratize the knowledge work being described above or if the rate of change and innovation within the artificial intelligence and machine learning spaces will continue for another decade before either saturation of exhaustion creates a great leveling effect within the technology.
Links and thoughts:
Yannic was on point this week, “[ML News] AI predicts race from X-Ray | Google kills HealthStreams | Boosting Search with MuZer”
I watched some of the WAN show with Linus and Luke this week, “Apple WON... but also Lost - WAN Show September 10, 2021”
You can check the Microsoft Developer AI Show this week, “AI Show Live - Episode 30 - Ramping up your custom NLP tasks with Verseagility”
Top 5 Tweets of the week:
Footnotes:
[1] https://www.brookings.edu/research/the-geography-of-ai/ make sure to download the report from the link on the left vs. just reading the page
[2] https://www.brookings.edu/wp-content/uploads/2021/08/AI-report_Executive-Summary.pdf
[3] https://research.google/teams/brain/
[4]
https://www.deepmind.com/
[7] https://www.linkedin.com/company/facebookai/about/
What’s next for The Lindahl Letter?
Week 35: Explainability in modern ML
Week 36: AIOps/MLOps: Consumption of AI Services vs. operations
Week 37: Reverse engineering GPT-2 or GPT-3
Week 38: Do most ML projects fail?
Week 39: Machine learning security
I’ll try to keep the what’s next list forward looking with at least five weeks of posts in planning or review. If you enjoyed reading this content, then please take a moment and share it with a friend.