OpenAI gets $1B from Microsoft

Microsoft investing in a company, rather than acquiring them, is not particularly common. OpenAI, however, has netted a $1 billion investment from the company to continue development initiatives in the artificial intelligence space.

OpenAI is a for-profit company, with investment from Elon Musk, Sam Altman and a number of others in the past. The company builds a number of products, such as Universe, a platform for measuring and training A.I inside games, websites or apps. 

Microsoft’s $1B investment, in my mind, is about exactly that banal use-case, rather than flashy research (which it will also do, of course) OpenAI will publish and invest in over the next few years. 

OpenAI is really interesting to me because that first product addresses a problem developers already blow tons of time on: data entry and banal app-based tasks. They abstracted real world problems into a playground for A.I. to learn from, where datasets just aren’t a problem.

Artificial intelligence is always heralded as this computer from the future thing, where it will solve our problems and cure disease, but where it’s going to shine are the boring, human-bridged problems of today. Banging data into Excel spreadsheets, typing into forms, scanning documents, whatever.

In the real world, often the dumb, hacked solution is easier than fixing a problem. I know of at least two companies that offer “modernizing” services for companies stuck on decades-old software, which isn’t anything more than writing a script to copy data from one page, scrape something from another, then ram it into a box on a new website. It’s cheaper than integrating their APIs, and means that you don’t have to actually fix anything!

A.I. Is perfect for these types of use cases, where it can figure out how to deal with this, and bridge the gap for the humans that might have otherwise done the job (or the hacked script that would break when that form slightly changes). The very same challenges are what make Google’s forthcoming service, that can automatically book rental cars or flights for you, very compelling.

OpenAI sits in an interesting space where it appears to be able to solve these boring problems, very pragmatically, while researching how A.I. will be used in the long haul. Microsoft’s investment has been spun as a crazy waste of money, but given that it runs a platform called Azure, where OpenAI will now spend a lot of money on A.I. Infrastructure, it doesn’t seem that unreasonable.

The best win for Microsoft out of this deal is that OpenAI will ‘jointly develop’ new tools for its Azure platform that make large-scale AI available for all... and those will be exclusive to Microsoft’s platform. Check, mate. 

Plus, sometimes you just want to get in the good books with the cool kid that’s doing the best research, so you can acquire them later. Given you’re figuring out how to completely reimagine the ways we interact with computers, it’s probably a decent bet to make, even if it takes a few decades to get there.

Tab Dump

Apple is in talks to buy Intel’s shuttered modem department
Today in wow, that isn’t surprising at all, Apple is reportedly looking at paying $1B or more to acquire Intel’s abandoned mobile efforts. If anyone else scoops this up, I’ll be incredibly surprised, and $1B is a bargain even if it was a mess, given there’s literally nobody else even in the game outside of Qualcomm. 

Big news: Slack is 33% faster, and shed a ton of RAM usage...while still being built on Electron.

Robinhood’s mobile trading app nets another $350M in investment
I always wished I could use this, as we don’t have many decent tools in Europe to access the markets. I make do with a crappy one called DeGiro, but it’s complicated and impossible to use. As savings rates have dwindled into irrelevance, millennials are learning all about the stock market, and Robinhood stands alone as something that actually helps make sense of it all.

Facebook Messenger for Kids had a bug that let kids do the one thing they shouldn’t be able to do

That thing? Talk to strangers that weren’t approved. The whole point of the app was to allow parents to properly manage who kids talk to, but a loophole allowed them to talk to anyone. It’s pretty clear that this company has a very serious problem: it has no idea how to build secure platforms.

Equifax will pay $650M or more to settle largest-ever databreach suit