Have you tried a virtual board, like Miro?

Not above commenter but I have. I have done OneNote, Miro, and several other canvas style apps. And for me it still just doesn’t click in my brain like writing it down.

It would take longer to train on intercepted Fable data, no?

Of course, but why be logical and think about the situation critically when you're pushing very hard for regulatory capture against competitors that give their weights away and provide services that are more reliable, offer a better value, and, this is the worst part, they're from CHINA.

Some of the accusations were going so far as to imply that they were outright routing your requests to anthropic and logging its/your response. It's kind of pathetic

Read the session file of the reviewer tomorrow. Sonnet should suffice.

Fighting against the weights is often a lost cause. For any model.

> The best agentic developers are now probably exceeding 100x, doing massive rewrites of codebases that would have taken years of engineering in days

I‘m not denying that you can get far with a port of a well tested codebase. But it’s a bit of a selective example, no? Porting a large framework to Rust is not something that’s making up a meaningful fraction of a developer’s time usually. It’s also a bit of a luxury IMO, and something you could have skipped entirely.

To add to this, for all the claims of 100x, can anyone cite an example of someone actually pushing 100x in a domain that isn’t “translate or port tool/library/whatever from language A to language B”? Because that’s literally what LLMs were invented for and that’s not a normal part of work.

I will fully admit we have examples of people converting Postgres to Rust, HL2 to wasm, etc. but that’s all porting. The code was there, we’re just translating it, warts and original-languagisms and all.

But even if you’re making a bunch of tiny tools and chaining them together, I can’t see more than a 2-3x sustained increase.

We also have artists that can make games now, and that 0-1 transition sure feels like 100x and maybe it is; but the problem of “starting feels impossible and the barrier is too high” to “I can now actually do something” is not where the 100x developer idea comes from either.

And that 0-1 change asymptotes really quickly to, again, maybe a sustained 2-3x once the artist understands the domain.

So, does the 100x person, who isn’t a novice and who isn’t literally translating existing code from A to B, exist? What are some real world examples?

Edit: I will concede one area: the investigation stage. E.g. “Does this code have an exploit?” I would argue that’s not generating new output (yet), but it is a great deal of work that is simplified and sped up. It’s also where most of my own use of AI speeds me up considerably as a developer.

I guess that’s why they call them Transformers

The trend to put acoustic panels or an imitation of their pattern on _everything_ is one of my pet peeves. They create some stroboscobe effect with my (not normal) vision. The result is that they mess with my 3D perception and break my brain a bit for that reason. I’d rather not have them put anywhere.

They are a fad, like many before, and not the best one due to copy&paste uniformity. Since its used so aggressively and so often, many places look weirdly similar like if many people bought exactly same Ikea furniture. Proper wood like oak can be very pretty in many forms but this is not.

In future it will clearly mark the era of (re)construction.

I’d recommend to instead write a de-slop skill that instructs to launch a sub agent with fresh context, and analyze for such phrases in the new commits, and remove those. Find -> fix just works better than preemptive instructions, in my experience.

And if you manage to do this automatically before committing, you’ve built the backpressure everybody is talking about.

Can you point to some examples? Also I wonder if this needs to be very model/harness specific? Like even model version, subversion.

And probably that should be run in different harness or with custom system prompt? Since they introduce quirks and glitches as well.

(somehow this motivated me to resurrect HN account)

I think you’re probably overthinking it. The same model will in my experience find errors in the same thing it just generated. I think reviewing is a totally different place in latent space, than implementing. Anyhow, it works well for me.

I often tell codex to launch a subagent without prior context to „remove BS phrases and make the prose sound more natural and higher readability“. That‘s usually enough to get better results.

I don't think that's opinions, but, rather, going wherever the wind blows.

Some people have a consensus-based mechanism for identifying if something is true or moral or fashionable. They basically believe whatever is the majority opinion, weighted for perceived social status or expertise or whatever else. I think this is normal for most non-technical people who don't have strong critical thinking skills themselves. So they outsource all the heavy thinking to the society at large.

There is no MacBook Pro with OLED (yet).

My mistake on tech; it’s a beautiful display. Alas, I speak from experience when it comes to the thermally-caused color shift. Hopefully it’ll be AppleCare covered.

I‘ve built this anti-k8s stance pre-LLMs, and just realized that, actually, agents should be pretty helpful in dealing with it? Is avoiding kubernetes still advisable for projects that will likely never use its full complexity, given how easy it is to maintain now?

Kube and helm are ideal for LLM usage, they make what is a fragile kind of thing into declarative, same as terraform for infra at the slightly lower level.