rablackburn

Excellent summary of the implications of LLM agents.

Personally I'd like it if we could all skip to the _end_ of Asimov's universe and bubble along together, but it seems like we're in for the whole ride these days.

> "It's just fancy autocomplete! You just set it up to look like a chat session and it's hallucinating a user to talk to"

> "Can we make the hallucination use excel?"

> "Yes, but --"

> "Then what's the difference between it and any of our other workers?"

You mean by the site founded by RCE-native AI agents?

> And if there’s one thing we’ve learned over the last decade, it’s the bitter lesson: when you have enough data and compute, learned representations usually beat carefully hand-tuned systems.

There are still holdouts!

Come back to me in a couple of decades when the trove of humanity's data has been pored over and drifted further out of sync with (verifiable) reality.

Hand-tuning is the only way to make progress when you've hit a domain's limits. Go deep and have fun.

> the rich's strategy of owning nothing

I think you may be coming into this conversation with a different definition of "rich" than most people.

I'm curious why you seemingly discount school as a "tight community structure". In many communities it's one of the only options left.

I daresay the label of the community is irrelevant, what matters is some other aspect effective ones share - and of course, the child in question (:

That was not an appeal to the authority of academic papers so much as the OP trying to give context for the information that has informed their position.

Your responses have been an appeal to tradition (“every generation thinks that”), and a dismissal of the information because of the reproducibility crisis.

Ie you are arguing that we (humans) struggle with discerning Truth, and therefore we are wrong, and everything is fine.

But taking the negative position is just as epistemologically flawed. Hence the OPs attempt to discuss the best data we can find.

> We are at the point in the technology curve with AI where every day someone figures out something new to do with them, meaning users will take any chip they can get, and use it productively.

Sure, that's an assertion.

But (with just as many citations), mine would be:

This boom is absolutely, 100%, fueled by the combined factors of: 1) employees outsourcing the cognitive load of their jobs to models that are, impressively "close", but not quite _as_ good as a well-trained human.

ie, we're replacing google with a fun, but terribly energy-wasteful (and _very often_ factually wrong) "make up an answer" tech.

and 2), AI "app developers" who are having fun with the previously "impossible" (*cf. https://xkcd.com/1425/) APIs of multi-modal natural language, and "didn't sci-fi warn us about this?" simulations.

Neither of which are good for productivity, if we measure productivity as "improving circumstances for the mutual commonwealth of all life". Which is the goal.

* oh, I _did_ use a citation after all.

It is an interesting article, but _far_ too sure of itself in all the wrong areas.

> It would be willfully ignorant to pretend that there's not an explosion of a novel and specific kind of stupidity

I 100% know what you mean, and largely agree, but you should check out the guidelines, specifically:

> Don't be curmudgeonly. Thoughtful criticism is fine, but please don't be rigidly or generically negative.

And like, the problem _is_ *bad*. A fun, on-going issue at work is trying to coordinate with a QA team who believe chatgpt can write css selectors for HTML elements that are not yet written.

That same QA team deeply care about the spirit of their work, and are motivated by, the _very_ relatable sentiment of, you DONT FUCKING BREAK USER SPACE.

Yeah, in the unbridled, chaotic, raging plasma that is our zeitgeist at the moment, I'm lucky enough to have people dedicating a significant portion of their life to trying to do quality assurance in the idiomatic, industry best-standard way. Blame the FUD, not my team.

I would put to you that the observation that they do not (yet) grok what, for lack of a more specific universally understood term we are calling, "AI" (or LLMs if you are Fancy. But of course none of these labels are quite right). People need time to observe, and learn. And people are busy with /* gestures around vaguely at everything /*.

So yes, we should acknowledge that long-winded trash PRs from AI are a new emergent problem, and yes, if we study the specific problem more closely we will almost certainly find ever more optimal approaches.

Writing off the issue as "stupidity" is mean. In both senses.

At least microplastics don't make you angry and violent that we can tell.

On the other hand, it's going to be around (relative to pre-emission levels) for a lot longer than the lead (paint gets chipped off and disposed of, we stopped using it in end-consumer products, etc)

I did read the post, but allow me to also recommend rambling when you’re remote.

As in, take time in your day to wander and roam. (I would go for a ~1hr hike in the mornings as my “commute”)

It gives you a sense of distinction from being home or “at work”. The routine cardio, and musings you have while walking make it well worth it.

> I have found that the old-fashioned “measure twice, cut once” approach is highly destructive. It was how I was trained, so walking away from it was scary.

To be fair it’s great advice when you’re dealing with atoms.

Mutable patterns of electrons, not so much (:

>> ...if it's the sole feature its users would care about?

The tag-line covers it pretty well I thought?

"git-bug is a standalone, distributed, offline-first issue management tool that embeds issues, comments, and more as objects in a git repository (not files!), enabling you to push and pull them to one or more remotes."

That tells you what the feature is - if you need/want a more technical overview you can still get from the `README` to `entity data model` in two clicks (Documentation > Data model).

I took it as capital-L Life is moving from high to low. As evolution continues Life seems to evolve ever higher -> lower/more-ordered organisms (as more complex organisms depend on the systems created by simpler organisms prior to themselves).

I am slightly blending the concept of entropy and complexity. But "ordered complexity" is how I imagine it.

>> I would love a good issue tracking system that is done entirely inside git

You might like git-bug:

https://github.com/git-bug/git-bug

Do you have any recommendations for where we could read more about the soviet originals/western derivatives? I’ve never heard that before (born after the collapse) and that sounds like a fascinating story.

> It's a bit like how vacuum cleaner specialists never use Dysons despite them having the most mindshare and overwhelming marketing budget. Regardless of vacuum cleaner specialists also just being normal people.

Yes, but Dyson still crushes the market because it has the mindshare with non-specialists. That’s OPs point, uninformed people (the vast majority) will be impacted more by the “ghibli-core” memetic moment, and have a greater impact on future events than specialists who appreciate the Gemini 2.5 release

Crazier things have happened.

Nice. SRS integration is my ultimate yak-shave when learning something new.

> I wish educators and educational institutions would make an attempt to incorporate SRS into classes. I think it would help a lot of students, especially for cumulative final exams.

Some of us try to. But when teaching middle-high school aged children the problem is almost always one of motivation and engagement rather than tooling and methodology.

You can take a horse to water, but you can't make it drink. Worst case scenario you try to enforce SRS activities and some students develop an aversion to an incredibly powerful tool.

My sweet spot was something like "here is an app called Anki, here is a deck of content we will cover in this class (tagged by week), try it out and see if its useful for you". Even then I was slightly conflicted because I've always wondered how much of the utility I've gotten from SRS came from sitting down and making the decks myself...

Oh neat, I do exactly this and thought it was just another instance of neurodivergent self-soothing behaviour. Turns out I’m just practicing my yoga ;)

I’m curious to know what your experience of changing to Warp has been like? Sounds like terminal was working fine for you - was the switch worth the time?

I signed up for the waitlist ages ago and finally got the announcement of Linux support in February but am still yet to try it. Mainly because I’ve had a particularly busy year and I can’t justify fiddling around with my stack just for fun. I have no appetite for the risk I may lose hours to fixing something that goes wrong.

> Why do I suspect the offence will always be ahead of the defence in these areas?

because destroying is easier than creating/entropy increases over time?

The only solution I can see is working on turning bad actors into good actors, or another way: positive reinforcement cycles.

No idea what that would look like with regard to LLMs though.

> Of whom? The person who made Quacker News?

Yes.

The “AI” didn’t do anything by itself, it was tasked to produce this. The author then made the decision to publish its output.

The fun-house mirror analogy breaks down because those being depicted aren’t opting in to this experience. It’s like a fun-house mirror that’s scraping people’s social media profile photos and mocking them for their looks.

It seems a bit mean-spirited as it picks specific commenters from threads to mock. Satire is one of those things where if you're not punching up it's just bullying.

> (...) and even Mario

Cool Japan would like a word with you ;)

And to quote the article itself:

>> (…) it’s a rare case of a kernel memory access check bug that had no security implications.

My favourite type of kernel bug.

This is one of the best job ads I’ve ever seen. Seamless segue into culture and “btw we’re hiring” at the end :)

This is truly a fantastic post-mortem: even an application-level developer like myself could follow it. Though I am in the middle of Rust in Action at the moment so I was primed for this type of content (:

And of course it’s always a pleasure seeing someone else who comments their code at such a high LoC ratio. Literate programming works!

>> You've been able to do exactly what you describe (i.e. putting "hello" squarely in the middle of a 600x600 yellow box) with CSS since its first incarnation, with e.g. the `valign` HTML attribute corresponding to the `vertical-align` CSS property on an `inline`, `inline-block` or `table-cell` box.

Notably not a `block` element like a div or p - which is what most people are going to use as a container.

They mention this in the article, and of course, the scientific method means we _want_ to have more than one suggestive meta-analysis before we declare any kind of "discovery" at all.

> Proteins (for example Glial Fibrillary acidic protein, GFAP) had previously been identified as potential biomarkers for dementia in smaller studies, but this new research was much larger and conducted over several years.

> I just can't imagine the scenario where you need to explain to someone who doesn't know how to modify their path why they need to add something to it.

Sounds like someone hasn’t had to train fresh graduate engineers for awhile ;)

> All security is by definition obscurity.

Can you expound on this? For example, if I add a 30 second lockout after failed authentication attempts I don’t see how that comes under any non-tortured definition of “obscurity”.