grumple

In January 2025 most of us didn’t even consider the frontier models good enough for regular usage.

This is how you get wrecked by malicious packages being published or other breaking bugs. Pin your versions.

If I were China, I would ban US models from being used by my government for the same reasons.

It’s true and I’m not. If the engineers can’t build software for your OS, your users won’t have software to use.

We’re also talking about MacOS, not iOS. Engineers make up a big segment of MacBook users. Apple is only 9.5% of the personal computing market.

Engineers and their companies are super users on whom the whole Apple ecosystem depends. If we can’t run our applications on Apple’s operating system, or they break things for our users, we will leave.

Take that lesson back to your company as well. It is easier than ever to switch between clouds and AI providers. It’s easier than ever to replace any software in house. This isn’t an idle threat; I’ve already moved off of several software platforms this year for other providers, and I’m probably doing the same for millions of dollars worth of business every year in the future as well. And my company isn’t immune to it either - our biggest customers are always looking for options to bring software in house. The more we disrupt them with “small bugs”, the more likely they are to leave.

The response from apple is insane, “please keep trying to reproduce this for us”. That is your job, apple engineer. As a manager of a large software org, a response like that from one of my engineers would drive me insane. Fix the damn problem or provide a timeline for addressing it. When you think the fix is released, then ask them if their issue was resolved.

I believe that is true, but more of it is still a liability. Good code tends to be simple. Bad code tends to be verbose, and verbose code tends to be bad. As someone who has been extremely prolific both before and after AI in terms of output, and who monitors the output of LLMs (including anthropic’s best models) I know that they are usually generating a bunch of junk. It will do what you ask but it’s largely terrible. Yes, you can mitigate that with a bunch of adversarial agents, but it’s still not great. Every line of code can potentially be a security hole or cause other problems.

Seems obvious to me that jobs which primarily require research and thought will be the first to go. The job of a good exec is basically: 1) research the market, 2) make presentations, 3) make high level decisions based on research and input from others and have leaders below you work on those. Ok, 4) is schmooze and maintain relationships but you have an org to do that too.

Fable is behind Opus (which is also much cheaper) at this point in benchmarking.

Surely all of china has the resources to match a single ai lab?

There are a plethora of qualified Americans with CS degrees sitting unemployed. Having worked with a lot of foreign tech workers - I’d take the Americans any day.

The H1B program today acts to undercut American workers.

If you’re looking at feeds and posts on LinkedIn hoping to get great content, you’re using it wrong. That is not what the site is for, that was just a sad attempt to emulate Facebook.

Here is how you use LinkedIn: you mark yourself as looking for work. You search for job postings and set up alerts. You apply for jobs. You respond to recruiters. If your friends share a post about a job you’re interested in, you apply. If you have a job opening, you post it.

And stop whining about the feed.

I’m using it a bunch. It saves a ton of time writing or reviewing code. It will catch things I won’t. But I’d express caution about the analysis or evaluation they do - LLMs will often confidently proclaim problems as solved or explain functionality and be wrong about it. Sometimes subtly, but sometimes just completely wrong. This is no different from humans, of course, except for the unabated confidence.

The ICC itself and its member states believe they have extraterritorial rights to interfere with legal practices elsewhere, which is sort of the whole point.