1) Yes.
2) No.
3) Yes.
Given how abysmal Copilot was/is, and how buggy their software has become this year, it's clear that they aren't reaping any benefits from AI usage internally.
Given how abysmal Copilot was/is, and how buggy their software has become this year, it's clear that they aren't reaping any benefits from AI usage internally.
most office workers seem to want at least the 20 usd per month plan
This can mean nothing who uses AI for programming, but is useful for day to day office work - and I belive it can sell very well
Office work is many fields in many businesses of all shapes and sizes in many industries. It doesn’t make sense to generalize.
Similarly, Google is improving Gemini and still dominates pretty much everywhere else from the phone to the car and on the web.
They don't have to have the best LLMs to finally deliver on the value they promised over a decade ago. Applications are always a much better moat, and it's not like anyone was going anywhere else anyway. I'm fully expecting Microsoft, Google, Apple, and Amazon to reap soon and buy up datacenters for pennies on the dollar. Market won't crash because China was smart enough to pause and maintain the hardware shortage.
> This can mean nothing who uses AI for programming
It doesn't matter because AI programming never got much buy-in outside of leaner teams working at startups. Startups will continue to migrate towards running their own local models. The corporate world does have some limited use for greenfield and misc tech debt, but will always be slowed down by the need for much larger scale human collaboration. The LLMs don't move the needle at all on productivity when the changes remain small (the parts people agreed on after all the teeth pulling) and the sprints and deadlines are still long.
MS must be losing mountains of money over this.