causal

Similar to the copy/paste issue I've noticed LLMs are pretty bad at distilling large documents into smaller documents without leaving out a ton of detail. Like maybe you have a super redundant doc. Give it to an LLM and it won't just deduplicate it, it will water the whole thing down.

If this is real it would be so easy to create a hype video that shows one of these things out in the wild handling unscripted situations with adversarial actors. But the fact we always see the same basic tasks in these videos makes me deeply suspicious that this is demoware. Bring in the Boston Robotics hockey stick guy for starters

Yeah tech companies have a weird fixation on using dystopian literature as their entire branding playbook

Haha thanks for this - I was irritated by the weird sort of empathy-without-empathizing the original had

I mean that's the narrative but can you provide any evidence this is happening with the Quiet Skies program?

Nobody is advocating for unnecessary checkboxing.

When 18-wheelers cross a bridge, it has to be engineered correctly. Unnecessary red-tape happens, but in no way would I want a bridge you "shoot from the hip" carrying my family alongside commercial trucks.

Also: kids click on links adult ignore without thinking. Our brains have built in filters for avoiding content we don't want; for kids everything is novel.

I enjoyed tab complete at first, but realized after a while that it was subtly changing what I meant to do. Too many suggestions forcing me to stop and read, interrupting the clarity I had on what needs done.

Agreed. If I'm typing code, it's probably because I know what I want to do, and suggestions popping up easily break my train of thought. I also hate the choice of tab - I use tab for...tabs.

There is probably a configurable way to request suggestions instead of having them automatically pop up - I should configure that.

Sounds like you're still using the terminal. I'm talking about https://marketplace.visualstudio.com/items?itemName=anthropi...

I do find the agents useful sometimes, but Cursor is most useful when it gives me visibility and granular control over what the agents are doing. Trying to follow a verbose narrative through a narrow chat pane is not it.

Yeah I've used staging/stashing similarly, but it feels like working around Cursor rather than with it

Addressing 2) first: That's good, I totally misunderstood then, and guess I'll need to try it to understand what's new since I thought that kind of tabbing had been there a while.

Back to 1): The type of bug I see most often is where conversation history seems incomplete, and I have trouble rolling back to or even finding a previous point that I am certain existed.

Git shares some features but I think Git was not made for the type of conversational rapid-prototyping LLMs enable. I don't want to be making commits every edit in some kind of parallel-git-state. Cursor's rollback and branching conversations make it easy to backup if a given chat goes down the wrong path. Reapply is tedious since it has to be done one edit at a time - would be nice if you could roll-forward.

I haven't put much thought into what else would be useful, but in general the most value I get from Cursor is simplifying the complex state of branching conversations.

Yeah I have to imagine it's because of user base, there's no real moat in the technology

I got into Cursor a little late, went really heavy on it, and see myself using it less and less as I go back to VSCode.

1) The most useful thing about Cursor was always state management of agent edits: Being able roll back to previous states after some edits with the click of a button, or reapply changes, and preview edits, etc. But weirdly, it seems like they never recognized this differentiator, and indeed it remains a bit buggy, and some crucial things (like mass-reapply after a rollback) never got implemented.

2) Adding autocomplete to the prompt box gives me suspicion they somehow still do not understand best practices in using AI to write code. It is more crucial than ever to be clear in your mind what you want to do in a codebase, so that you can recognize when AI is deviating from that path. Giving the LLM more and earlier opportunities to create deviation is a terrible idea.

3) Claude Code was fine in CLI and has a nearly-identical extension pane now too. For the same price, I seem to get just as much usage, in addition to a Claude subscription.

I think Cursor will lose because models were never their advantage and they do not seem to really be thought leaders on LLM-driven software development.

IDK if the site is being hugged to death but I can only load the first video. Even in just one viewing there were noticeable artifacts, so my impression is that Veo is still in the lead here.

Don't think I've ever preferred Opus to Sonnet

I'd argue that dressing an ad up as an academic paper is obfuscation

Yeah, especially since many of those are just poor targets after being out so long / contaminating too much.

That has got to be the most benchmarks I've ever seen posted with an announcement. Kudos for not just cherrypicking a favorable set.

He might be right about intelligence becoming the new currency in a world where intelligence becomes fungible.

Lots of assumptions about the path to get there, though.

And interesting that he's measuring intelligence in energy terms.

I like that people are experimenting with HTMX, so please keep it up. But I don't think MESH is one I would pick up for my projects:

- For complex apps, I love JSX for its ability to distill complexity into simple, discrete components.

- For simple apps, I love HTMX ability to make plain the relationship between raw HTML and API response.

In the MESH examples I see a lot of verbose code that doesn't seem to achieve the simplicity of JSX nor the transparency of HTMX. That said, I certainly don't grok MESH yet so perhaps in a project it would click for me.

I think that is the bigger difference, I agree, but I think the format and media involved still have an effect on the type of discussion that takes place

I wonder how much of the difference is just because HN doesn't allow you to post / display images inline like Reddit does, forcing users to engage with the content (written or video) rather than just reflexively upvote low-effort content.

This comment ironically at the top

For me it was very sensitive to even one red pixel from a sign not being selected

Started laughing out loud at the recursively smaller stop sign squares

1) Data proximity (if you have a lot of data, egress fees add up)

2) Hardware optimization (the exact GPU you want may not always be available for some providers)

3) Not subject to price changes

4) Not subject to sudden Terms of Use changes

5) Know exactly who is responsible if something isn't working.

6) Sense of pride and accomplishment + Heating in the winter

I meant that the authors of the study should have gone much deeper, and WaPo should not have published such a lazy study.

Not true at all. You can take a vector for a given embedding and compare it to other things in that area of latent space to get a sense for how it is categorized by the model. You can even do this layer by layer to see how the model evolves its understanding.