Microsoft's head of AI doesn't understand why people don't like AI, and I don't understand why he doesn't understand because it's pretty obvious

Microsoft AI CEO Mustafa Suleyman says it's "mindblowing" that people aren't more impressed with generative AI tools.

"Jeez there so many cynics!" wrote Suleyman on X this week. "It cracks me up when I hear people call AI underwhelming. I grew up playing Snake on a Nokia phone! The fact that people are unimpressed that we can have a fluent conversation with a super smart AI that can generate any image/video is mindblowing to me."

The comment comes as the public reacts to Microsoft pushing its new "agentic" services, referring to AI agents that can perform tasks beyond chatting and generating images. Windows head Pavan Davuluri said earlier this month that Microsoft's OS is "evolving into an agentic OS," and I can't say I'm excited about the prospect.

Suleyman's inability to comprehend that lack of enthusiasm can easily be resolved with a classic technique from ancient philosophy, or maybe it was an ad for taco shells, in which one accepts two different things at the same time.

In this case, I think that generative AI models are very novel and that it's absolutely appropriate to be curious about the ways they mimic human creativity and reason, and without contradicting myself at all, I also think that they've been recklessly jammed into commercial software despite not actually being capable of the things tech companies claim they are.

In Suleyman's brief tweet alone there are two obvious falsehoods: AI chatbots are not "smart" and they cannot "generate any image/video." They can do lots of interesting and peculiar things, but fail even at the simple tasks Microsoft shows them doing in its ads.

For instance, The Verge recently attempted to recreate a moment in one of Microsoft's commercials in which Copilot correctly identified the location of a cave from a photo. In The Verge's test, the Windows chatbot repeatedly told them the location of the image in the Windows file system rather than its location on the planet (it'd be a good gag if chatbots could have a sense of humor), and when it did provide a geographical answer, it was wrong. The reporter also found that they could get Copilot to claim that the cave is in New Jersey simply by adding "new-jersey" to the file name. (It's in Mexico.)

That's just one of countless examples of AI failing to do anything but waste our time. I Googled "Black Ops 7" last week, and Google's AI search results told me it was a made-up game that doesn't exist. Helpful!

And I haven't even mentioned the scraping of copyrighted material that makes these bots possible in the first place, the ugly AI-generated art showing up in videogames and other media, the dangers LLMs pose to vulnerable people, the techno-soothsayers claiming that these busted chatbots can do all of our jobs, and the enormous resource investment going toward AI data centers—all to rapidly commercialize poorly understood technology that doesn't actually do the things we're told it does.

Is it really so hard for the tech industry to understand why not everyone is bowing before their new gods?

AI and machine learning may in fact transform the world, but there's no reason to assume that the transformation will be a good thing unless we actively try to make it a good thing. So far, tech companies have given no indication that they care about anything besides the pursuit of profit. I don't think that we're the cynical ones here.

Best gaming rigs 2025

1. Best gaming laptop: Razer Blade 16

2. Best gaming PC: HP Omen 35L

3. Best handheld gaming PC: Lenovo Legion Go S SteamOS ed.

4. Best mini PC: Minisforum AtomMan G7 PT

5. Best VR headset: Meta Quest 3

👉Check out our list of guides👈