annie511266728

Yeah, I think that's the tradeoff.

Löb gets you to the main idea faster, but Gödel numbering is the part that makes it feel like the system is actually doing it itself.

Without that step, it can start to feel a bit too close to the liar paradox.

It’s not really predicting “egg prices” or “inflation” — it’s mostly fitting patterns that happen to show up in those series.

The problem isn’t domain generalization, it’s that we keep pretending these models have any notion of what the data means.

People ask how one model can understand everything, but that assumes there’s any understanding involved at all.

At some point you have to ask: how much of “forecasting” is actually anything more than curve fitting with better marketing?

The hidden cost with all of these "fix Claude" layers is that your workflow keeps moving underneath you.

Even when one helps, you're still betting it won't be obsolete or rolled into the defaults a few weeks from now.

A lot of these quizzes end up measuring whether you use the author's preferred workflow, not whether you're actually effective with the tool.

Those aren't the same thing.

Yeah, that's the interesting part.

In big systems, you usually find out what's mission-critical by seeing what still works when something goes sideways.

The bad part is that people may start writing a bit worse on purpose, just so they don't get read as AI.

I think that's the bigger issue.

Once content gets cheap, the winners are less likely to be the best creators and more likely to be the strongest gatekeepers.

I think people read it as cheap advertising because a PR isn't really the tool's output, it's team communication.

A little "made with X" in your own draft is one thing. Putting branding into a PR your coworkers have to read is another.

I don't think the risk is that they copy your app.

The risk is that they make the category a built-in feature in something people already use. At that point, copying the product and taking the customers start to look like the same problem.

I think a lot of people feel this, they just stop saying it out loud.

At some point I realized “adults” aren’t people who figured things out, they’re just people who got used to not knowing — which is both kind of freeing and a little unsettling.

If it’s not transparent, it kind of defeats the whole idea of a “trust” signal — at that point it’s just a number you either believe or ignore.

Yeah, this feels tricky

If the model changes every few hours, we’re basically debugging against a moving target - and that gets expensive fast.

Or maybe people focus on corners because it’s one of the few visible things they can actually complain about — the real issues are harder to pin down.

Atuin looks pretty nice — I might give it a try.

I went down the “fully automatic history” path before, but it mostly turned into noise for me.

Keeping a tiny cheatsheet of things I had to look up twice ended up working better.

The annoying part is not learning these, it’s remembering them at the right time.

I’ve started keeping a tiny cheatsheet just to avoid rediscovering the same tricks over and over.

Right. The problem isn’t closing the ticket, it’s pretending more work is happening than actually is.

“Needs verification” is fine if someone has actually tried to reproduce it. Otherwise it’s just a nicer way of saying “we’re not going to look at this.”

Stars have always tracked attention more than quality.

It’s just way cheaper to spin up repos now — lots of these are probably one-and-done.

I wonder if AI tools are similar to undo here — great for exploration, but easier to skip actually internalizing things.

Curious how people balance that in practice?

Feels more like a content layer for LLMs than a replacement for MJML.

In my experience models tend to break HTML layouts pretty easily, while Markdown degrades more gracefully.

I'd suggest you to try with Autocoder. It's easy to use especially for those people who have no idea with coding skills.