ChrisLTD

"Our AI makes your workers X times more productive"

"Our AI takes the drudgery out of X, Y, and Z"

The datacenter antipathy is odd, but then again so are the AI corps' marketing strategy of making everyone afraid for their jobs and the future of the species.

If Linux doesn't qualify as open, then what does?

and if you're an author that lives outside the United States?

If the US implements export controls and China doesn't, it would kill the market for OpenAI and Anthropic.

Why would this law make things worse?

What would your proposal be for fixing what you’ve identified as the underlying problem?

You have to start somewhere, no? We have laws against stealing and murder and folks don’t usually go around saying they should be removed from the books because some people still steal and commit murder.

So why do you recycle? Based on your argument, you're just wasting your own time or "shooting yourself in the foot".

I've been using Parakeet Multilanguage locally, and it's super snappy

It's illegal to build dense cities like Manhattan in most of the United States. And while most people want to live in a Manhattan'esque area, plenty (like me) do.

If anything the first movers would theoretically be the first to destroy their own labor markets and economies.

The Supreme Court gave Trump immunity, stymieing Democratic attempts at holding him accountable for crimes he was accused of in his first presidency. Holding him accountable this time would require some workaround for the Supreme Court.

Yeah... Sonnet becomes the new cheap model, and some Fable class model becomes the more expensive/better one.

Their valuations don't make sense as just programming tools, period. Forget about if they are still human driven.

It's decent at rote coding tasks, but I haven't seen these things be reliable enough outside of that specific task to make the claim that it can do the work of any information worker.

What would it mean to be incompatible with the other 5.x models?

Startups and early stage businesses have always had less intelligence (when intelligence would be measured by the number and quality of their employees) than larger businesses. That hasn't stopped them from succeeding before.

If it's a new generation why isn't it GPT-6?

Seems like a structural problem that the U.S. elected who they did twice.

You're over-indexing on the (erroneous) idea that my example meant that the program would be a single if statement.

A lot of solutions require way more code to implement than to describe the necessary properties of.

That's true to an extent, the additional code often define the emergent and undiscovered properties of a system.

I can be 99.99999999% certain when I write an if statement like "if (x > 1) do y" that the compiler will turn that into the equivalent machine code. So, yes, unless I hit some crazy performance bottleneck, I'm not concerned about reviewing the machine code.

However, LLM outputs change with slight re-wording of prompts and with each new model release. I could hand write a test that says if x > 1 make sure y happens, but then what productivity was gained?

How can the code no longer matter? It literally is the logic (not to mention performance, and reliability) of the software.

"project requirements are deterministically met" makes it sound so easy

These aren’t bugs in the traditional sense. They built the animation system to work like this, and replaced the old system that didn’t produce these psychedelic transition states.

This 100%. Maybe I have a prompting skill issue, but without my guidance Opus (and now Fable) writes some gnarly stuff with tons of small bugs, and weird design decisions.

I agree with most of what you're saying, but it's also funny that front-end developers are catching strays in your conclusion. As long as human beings are interfacing with software, there's going to be a lot of judgement and nuance necessary for building good UIs. Data structures, back-end infra, can all be alien and still work. But your UI can't be alien.

Even the early versions of AI autocomplete tools like Tabnine and the original Copilot could autocomplete entire functions, so I think you might be strawmanning a bit.

If it’s not larger, it’d be tough to justify the massive price increase for using it.

I’m not sure the parent comment is talking about running the open weight models on your own hardware

more like 2 drinks per meal in Mad Men