And now we see the beginning of how even local LLMs will be turned against their users -- by persuading agents to advertise to them.

I don't think that's what you're intending here, but it's the next logical step. Agents are on the Internet, and they represent an opportunity to reach their humans.

I miss reading things written by humans.

Digital: https://libbyapp.com

Analog: https://bookshop.org

I am starting to see so much consistency in the "it's not AI, it's overhiring" commentary that it's actually starting to feel like a narrative constructed to allay concerns about AI impacts. At this point it's a "pandemic overhire correction" that the industry has been doing for two years, and is accelerating.

Yea, I don't know how long they're planning to milk the "pandemic overhiring" excuse. Ten years? In 2030, we'll still be seeing headlines like "Company X lays off another 10,000 workers due to overhearing ten years ago..."

That's a huge gap for llama.cpp server - any idea why?

Best guess is it's native mode. The function calling template is just broken for Nemo.

I did go with an extreme example in the post (but true). Other deltas are smaller but still statistically significant. 30 pt swing between llamserver prompt vs ollama, 4-5pt swing between llamafile and llamaserver prompt.

What chances do the vast majority of those graduates have to shape what's happening? That happens at exec level at the largest companies. Everyone else gets to produce or consume what they decide on.

Exactly. I work at Google and I’m relatively high level. And I’ve got zero input into AI being shoved into every surface. What influence will these grads have?

Did your generation think you would get to have any influence?

The same generation that is using those companies for everything.

Wanting the benefits but not its downsides.

May have been better as: "I like stawberrries, and walking to the car wash to clean my car."

You're absolutely right! My mistake. I should have said: "I like starwberries, and driving to the car wash to clean my car."

The majority of users will see the convenient answer right in front of them and stop, because their question has been answered. We've seen it again and again across industries, an an accelerating cycle: make it easy, and the users will usually do what you want.

Some percentage of developers before AI were unable to code fizzbuzz. Some significantly higher percentage of them are not able to do so now.

Saying there have always been bad developers doesn't change that there's a higher ratio of them now.

No stats to back this up. Just interviews I've done recently and historically.

This was nothing like investigative journalism; it's just LLM spew. It could have been written in a handful of paragraphs.

This isn't even well-slopped slop.