Every week, there is a new announcement about which company will win the AI race. This week, all the rage-baiters are posting “Claude is dead” posts and “Anthropic will lose all of its money; KIMI is crushing the metrics!”
Trying to keep on top of who is winning is spinning me around so fast I am getting dizzy.
I want to write a sentence like, “but it is not the model companies who will win, it will be companies who… “ and then I stopped because that is one of those LLM patterns that now triggers me so hard I can’t even complete the sentence.
Now that the spinning and self-loathing has subsided, I want to register some predictions.
This is likely the first “next big thing” race that is going to be won by incumbents.
Unfortunately for the investors in the big models and God knows whatever, this was always going to go to Google or Apple.
Google has been going quietly into deeper suites of tools at the software level, which is where they have a lot of experience and one or two successes outside of search. You might go “ohhh sick buuuurn” and reach for the ointment, but they are not very good at building usable things. I thought the name Google Plus was appropriate for their social software, because everything you wanted to do with it took Plus One clicks compared to a competitor product. That is still true today. They have a giant search playground, a lot of capital, and a lot of patience. They are also innovating in all of the right places instead of greedily trying to suck up their partners’ businesses 100m verticals at a time.
Apple, on the other hand, is making a completely different bet. They are going for the junk. They want to grab the whole LLM market by the short and curlies and squeeeeeze. Their goal is to put a Mac Mini or Mac Studio in every cubicle, bring on-prem LLM hotness to the market, and make the “90% as good, but free” models accessible to businesses for all their fidgety bits and inferences.
These are both good strategies, and oddly, they will work together. Much like the duopoly for the touchscreen device in your pants, there is a very good chance these two companies, one going high, one going low, are going to do this whackadoo pincer move on the startups who are sitting in the middle singing the “I can spell LLM” song over and over, and suddenly the lights are going to get turned out.
Usually the arrival of a “next big thing” heralds a new bajillion-dollar company. I think this is going to be one of the first times that a “next big thing” is going to go to the incumbents.
Does this work well with what is paying your bills today?
If not, then you have been warned.
See you next week!
