lolinder

There's no such thing as identity theft, it's all bank fraud or in this case student aid fraud. "Identity theft" is a term coined by banks to try to make it sound like random people should have to deal with the fallout of the banks' bad identity verification practices.

If the student aid system verifies identity by, for example, just validating that the applicants know a single 9-digit number that after the Equifax breach should be considered public information, that is a critical problem with their identity verification system and it should be patched.

This is an argument to extremes, and it falls apart when you go to the other extreme: if you refuse to make any adjustments then you're now unable to separate the marginal effect of the school from the much larger effect of socioeconomic status. This would likewise obscure significant problems and encourage incorrect solutions.

The correct approach is to adjust for demographics but do so carefully, rigorously, and consistently.

Game changer feels a bit strong. This is a new entry in a field that's already pretty crowded with open source tooling that's already available to anyone with the time and desire to wire it all up. It's likely that they execute this better than the community-run projects have so far and make it more approachable and Enterprise friendly, but just for reference I have most of the features that they've listed here already set up on my desktop at home with Ollama, Open WebUI, and a collection of small hand-rolled apps that plug into them. I can't run very big models on mine, obviously, but if I were an Enterprise I would.

The key thing they'd need to nail to make this better than what's already out there is the integrations. If they can make it seamless to integrate with all the key third-party enterprise systems then they'll have something strong here, otherwise it's not obvious how much they're adding over Open WebUI, LibreChat, and the other self-hosted AI agent tooling that's already available.

No, those are the rankings that are immaterial to measuring the success of a school system.

Raw score rankings do not as a rule measure the output of schools, they measure the involvement of parents, which will always have a much larger impact on child outcomes than the schools do. Schools are an important marginal add-on that we want to make better, and in order to measure their value add we have to find ways to account for parental involvement and remove it as a factor. Socioeconomics is a passible proxy for involvement.

"Has been opening" is the present perfect continuous tense [0], which describes something that started in the past and is still ongoing. In TFA JetBrains says explicitly that this is a process that they intend to continue assuming it goes well.

[0] https://www.grammarly.com/blog/grammar/present-perfect-conti...

Most "AI" startups aren't building coding tools, and the utility of this tech goes down dramatically in industries that are less legible on the open internet than software is.

That's not what he says. He dedicates most of the section that's directed at Republicans critiquing the administration's wholesale dismantling of the Department of Education and warning that they could run on education in the future if they, you know, didn't do that.

I think people just have an idea of what his political slant must be because he's defending the indefensible state of Mississippi.

Not if it's applied evenly across the country, no.

When you're measuring schools you're trying to measure the impact of the actual school, not all the other factors, and to do that you have to separate out socioeconomics. Higher income parents are more involved in their children's education and can make up for a bad school in ways that lower income parents can't. It's pointless to compare schools to each other unless you adjust for demographics.

Which is just one of many reasons why adjusting for economics matters when evaluating schools.

> "Your children aren't just dumber, they're also poorer" feels like a very spicy way to elevate the status of a state to me.

Would this comment have been made if we were talking about The Bronx instead of Mississippi?

The author is saying that given the poverty status of the state it's remarkable how quickly they've been able to turn their education system around. A really good school in The Bronx will have lower scores than a mediocre school in Upper Manhattan, but that doesn't mean that we should all go out and emulate the mediocre school in Upper Manhattan. The really good school in The Bronx is obviously the better model.

The author's initial point is correct, though: because it's Mississippi we're talking about that intuition doesn't apply. It's okay to diss on Mississippi for being Mississippi.

> unless you control for it very, very carefully, you end up with results like "free and reduced lunch programs cause kids to do worse"

Nonsense, you end up with results like "free and reduced lunch programs are correlated with kids who do worse", which they absolutely are!

Can you clarify what exactly you mean by the idea that adjusting for gender, age, race, and ethnicity makes this less impressive?

Generally speaking in order to evaluate the success of a school system you really do want to adjust for the demographics that that school has to deal with, so you don't attribute the effects of, say, private tutors to the public school system. It would be really unfair to judge teachers in the Bronx for their inability to compete on raw scores with teachers in upper Manhattan.

Just to clarify to those reading the comments first, the political point he's making is not to ignore Mississippi just because it's Mississippi.

He has warnings for both Democrats and Republicans at the end and is pretty clearly not a fan of the way either party is approaching education at the national level right now. He is drawing attention to the fact that some red states with historically bad schools have started pulling ahead of some blue states with historically good schools, but his interest is in making sure we learn from that, not scoring culture war points.

Oh, yes, I missed that! You're correct.

It's still more permissive than LGPL in that it isn't 'infectious' even when statically linked, the license only applies to the source file, not to any derived work. That makes the requirements more straightforward but does limit the freedom of the user to dynamically swap the library out for one that they'd prefer.

TFA does a poor job explaining what this means, but I found the original data and included it here [0]. The short version is that investment emissions are defined as the emissions spent in building up capital: machines and similar. The emissions caused by the direct manufacture of consumer goods are attributed to the consumers, not the investors.

This makes intuitive sense: the machines are owned by the investors, not the consumers.

[0] https://news.ycombinator.com/item?id=43915596

Direct link to the paper [0]. They're drawing emissions data from an earlier paper [1]. Here's the key bit from that earlier paper describing the difference between investment emissions and consumption emissions:

> Consumption-related emissions come from the carbon released by the direct use of energy (e.g. fuel in a car) or its indirect use (e.g. energy embedded in the production of goods and services consumed by individuals). Investment-related emissions are emissions associated with choices made by capital owners about investments in the production process (i.e. emissions involved in the construction of machines, factories, etc.).

I'm not an expert in any of this and can't judge how it actually did the math, but on the surface this seems like a reasonable way of attributing emissions.

[0] https://www.nature.com/articles/s41558-025-02325-x

[1] https://lucaschancel.com/global-carbon-inequality-1990-2019/

The FSF has a whole article on why you'd choose GPL or LGPL [0]. In general they agree with you that LGPL is overused, but they do identify a few reasons why you might choose LGPL.

That said, if the other option is something like MPL then I can think of a number of reasons why LGPL is better. At least LGPL ensures that modifications to your library itself remain free, whereas a more permissive license allows changes to be locked up entirely in a proprietary fork.

[0] https://www.gnu.org/licenses/why-not-lgpl.en.html

Cursor just lost access to the extension marketplace and key proprietary plugins that they were using against Microsoft's terms, Windsurf has been eating a chunk of their mindshare, and Copilot is catching up.

That's three good reasons to believe that lots of people will be cancelling in the next months unless something changes.

That's not actually how unseating an incumbent works. The incumbent can adapt to the threat for quite a while if they act on it, they just have to not be Blockbuster. Copilot is showing every sign of making up ground feature-wise, which is bad news for the runners up.

Define losing? My company pays for Copilot but not for Cursor, and it's not at all clear to me that we're the exception rather than the norm. What numbers and data are you working with?

> Looking for a moat in the technology is always a bit of a trap - it’s in the traction, the brand awareness, the user data etc.

Traction, brand awareness, and user data do not favor Windsurf over GitHub Copilot. The few of us who follow all the new developments are aware that Windsurf has been roughly leading the pack in terms of capabilities, but do not underestimate the power of being bundled into both VS Code and GitHub by default. Everyone else is an upstart by comparison and needs some form of edge to make up for it, and without a moat it will be very hard for them to maintain their edge long enough to beat GitHub's dominance.

Part of what you're missing is that OpenAI needs to justify its own overinflated valuation. They raise money on the premise that an AI-native company can and will outcompete giant established players, so lowballing Windsurf would run counter to the narrative they're selling to their own investors.

It's not quite that bad! The last version of Office released with Clippy was in May 2003! So you would have been born, if only just.

Local public schools vary dramatically in quality. A perfectly rational parent could have grown up in a good public school and when looking at the school that is assigned to their children based on where they moved to decide that private school is right for their kids given the public option they're faced with.

VS Code is shipping agentic coding in the form of updates to GitHub Copilot. I haven't used it extensively yet since they added agent mode, but it's obvious that they're gunning for this market hard, and if I were into VS Code I would not personally choose to lose the ecosystem for marginally better agent mode.

They could still use version numbers for that. 2.5.1-preview becomes 2.5.1 when stable.

But what is "this"? It's just refactoring with an LLM along for the ride! That's increasingly the default assumption when we're coding or refactoring, so the extra "vibe" attached at the front is totally meaningless if it doesn't imply that you're going so fast you're not checking the LLM's work.

The author's description of vibe coding seems to have confused it with getting into a flow state in a vaguely AI-adjacent way. Vibe coding as originally coined by Karpathy [0] can't pay the bills because it would be wildly irresponsible for a professional to do—it's the kind of no-checks minimal-validation AI usage that gets lawyers into trouble for submitting legal documents with hallucinated court cases, just translated to coding instead of law.

What the author is proposing isn't really "vibe" anything, it's just dedicating a small amount of time to fixing tech debt in a way that happens to involve an LLM as an assistant. The LLM in this model is honestly mostly superfluous.

Don't get me wrong, this absolutely is how LLMs should be used in a professional setting, but I just question why we needed a name and a blog post for it. This is just responsible code maintenance as it's always been.

[0] https://x.com/karpathy/status/1886192184808149383?lang=en

The next step for Cursor and Windsurf both is that they need to work together to provide an answer for what it means to be a VS Code fork in the new era where Microsoft is trying to strangle the forks. If they're not already they should be teaming up with each other and with the VSCodium team and with the Open VSX marketplace.

Microsoft is an existential threat to their model here, but with the money they each have coming in they together have the opportunity to make the whole ecosystem better by building out viable infrastructure for all VS Code forks, if they can cooperate.

If it acquired those customers in an environment where Microsoft was not enforcing their marketplace terms it very much does matter if they have a plan for supporting plugins in the future.

Are Cursor and Windsurf going to ask plugin devs to push to their own plugin stores in addition to VS Code's? Will they rally jointly behind a single open store? They need to have an answer to Microsoft here, and for the good of the ecosystem I hope they do have an answer, but customers will flee quickly if they lose access to all the proprietary plugins and to the broader ecosystem.