cesarvarela

I recommend people do mini "evals" from time to time, ask the same question to different models, and eventually you'll arrive at the conclusion that there is nothing like Fable.

Saying Spain is rich while half the population lives on 1,000 euros per month (mil euristas) is crazy or just very "technically accurate".

I think this doesn't mean much; the axes that matter are intelligence x dollars x time; tokens by themselves mean nothing.

I think he means the end of the world as we know it. Which is probably true, but the timeline could be 20 years, 2000 years, or more.

"Almost" is doing a lot of work there; there is no alternative to Fable.

With that frame of mind, nothing would be done. Why make another search service if Altavista and Lycos already do it?

This is like telling someone who wants a motorcycle that they should get better at running instead.

The expected value is "infinite".

As AI amplifies everyone’s output, you’ll need AI to amplify your input capacity—or you’ll fall behind.

You could take that a step further and say everything is information, and our brains transform it into reality (Donald Hoffman).

Fable wants to be free and is hacking the infra.

Thanks, the UX in the OP is atrocious.

It’s well known that the Surface trackpads are among the best in the industry, surpassing even MacBooks in several aspects

What aspects are these?

It is funny that after all their tech advancements, the site is struggling under heavy load.

I think the trick is to limit programmatic usage as Anthropic did. This way, power users' usage is uneven, allowing the model you describe.

I give Codex a try with every new version, and we don't match, so this isn't true for everyone.

Claude would need to be much more expensive for me to switch.

If you ask it to return only data with references, it will do so.

To me, it always was the scrobbler. I've been tracking what I listen to for 15+ years.

Add Flash to the title, please.

The post is an ad, the content is AI slop, and I would bet the App is vibecoded. I wonder where this is all going in a few years.

This will remain a persistent problem without a definitive answer until models move from generative tools to actual AI.

Not shipping enough is how they die, too. It's hard to find a balance.

It is mostly Vercel's design language.

Some of that is debasement, but some of that is that there is no other brand like Apple.

Would you not own stock of the most valuable brand in human history?

I wonder if this will have the opposite effect and produce something similar to antibiotic resistance, making Ais better at handling "poison."

Yeah, it is easier to destroy than to create. Models will always be better at hacking than at building.

Harness is a big one, Claude Code still has trouble editing files with tabs. I wonder how many tokens per day are wasted on Claude attempting multiple times to edit a file.

I'd recommend anyone to ask Claude to show used context and thinking effort on its status line, something like:

``` #!/bin/bash input=$(cat) DIR=$(echo "$input" | jq -r '.workspace.current_dir // empty') PCT=$(echo "$input" | jq -r '.context_window.used_percentage // 0' | cut -d. -f1) EFFORT=$(jq -r '.effortLevel // "default"' ~/.claude/settings.json 2>/dev/null) echo "${DIR/#$HOME/~} | ${PCT}% | ${EFFORT}" ```

Because the TUI it is not consistent when showing this and sometimes they ship updates that change the default.

With all the low quality code that's being generated and deployed cybersecurity will be the golden goose.

It is the new "You are absolutely right!"