iiinova — where your digital self keeps evolving

iiinova is part of the Atmosphere — an open, public network where you own your identity.

ben

@benjamineskola.hachyderm.io.ap.brid.gy

17 followers 0 following 676 posts

Software engineer, socialist. London, but from up north. Learning Finnish. #nobot Anti-AI, pro-human. 🌉 bridged from ⁂ hachyderm.io/@benjamineskola, follow @ap.brid.gy to interact

Posts Replies Media

ben @benjamineskola.hachyderm.io.ap.brid.gy · 6h

@bodil a naive prime factoring algorithm would just skip all even numbers meaning 2 is actually the safest prime.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 8h

@wess Yes. It’s hard to deny there will be an impact — but I don’t think we have to grant it the inevitability or agency that ‘driven’ implies.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 10h

@quixoticgeek I think I’d go for something like “we did discuss [alternative / mitigation] at the time but we decided against it”. Let them fill in the specifics of ‘we’.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 19/10/2026

@alter_kaker Yes; you can’t just apply the same tactics in pursuit of a different goal, the choice of tactics is part of the goal.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 19/10/2026

[slop and the people who love it] anyway. the 'vibecoding' tag there was not ideal, because it was used not only for projects that use LLMs, but also both for defences of and critiques of LLM usage. i feel like those are different things that one might legitimately want to filter differently […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 19/10/2026

[slop and the people who love it] but also: the LLM boosters on lobste.rs are complaining that the site is overrun by LLM haters and they're constantly under attack; etc which is funny because from my perspective it's overrun by boosters and it's nearly impossible to escape vibecoded rubbish […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 19/10/2026

[slop and the people who love it] what's striking me particularly about the discussions on lobste.rs are how defensive the pro-LLM people are you're winning! your garbage is everywhere! it's inescapable! and yet you're all throwing your toys out of your prams because some people want to label […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 16/10/2026

@gedankenstuecke interesting that it’s part of a union contract, because a few years back they fired all their staff when the staff tried to unionise. It’s a good change but I have my doubts about the publication still.

Reposted by ben

Ben Ramsey @ramsey.dev · 16/10/2026

“When reviewing PRs generated by AI, the cognitive effort never decreases.” IOW, you can’t build trust over time with the person submitting those PRs, and every time you review their code, you have to apply the same high level of critical thinking to assess its quality. (Paraphrase of a […]

Original post on phpc.social

ben @benjamineskola.hachyderm.io.ap.brid.gy · 16/10/2026

LLM advocates love to claim constantly that LLMs produce better code than they or anyone else could be expected to write, and then you see the code and it's a hideous pile of garbage.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 16/10/2026

it's not even that it's wrong to say 'most developers who think they are above average are not'. (by definition most developers are not above average.) but the conclusion to that doesn't have to be 'so just let the LLM do it'.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 16/10/2026

it's a little bit embarrassing to see someone who seemed to be presenting themselves as a 'sensible' voice on LLMs repeat the same tired nonsense like 'left behind'. (see also simon willison)

mastodon.social

ben @benjamineskola.hachyderm.io.ap.brid.gy · 15/10/2026

it helps that in most cases i seem to use a subset of the features and so i can just focus on that.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 15/10/2026

I seem to be on a minor streak of reimplementing shell tools that I use regularly but have started accepting slop. Turns out none of them were irreplaceable. 🤷

ben @benjamineskola.hachyderm.io.ap.brid.gy · 14/10/2026

@glyph I’d be interested to read something about it if you ever figure out a good answer. Part of me is increasingly suspicious of utilitarian reasoning these days anyway; and the EA/longtermist versions are part of the reason why. (And yet some ethical tradeoffs do seem justified, so I need to […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 14/10/2026

Looking forwards to having to smuggle her next book into the country. www.theguardian.com/uk-news/2026/oc…

ben @benjamineskola.hachyderm.io.ap.brid.gy · 14/10/2026

@zzt checking a codebase for slop earlier I found a commit titled something like “ensure single blank line between changelog entries” and the entire diff was deleting one extra blank line. Assisted by Claude Opus 9000 or whatever

ben @benjamineskola.hachyderm.io.ap.brid.gy · 14/10/2026

Think I’m going to have to mute the word ‘bot’.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 14/10/2026

Seeing multiple projects starting up (or restarting) around porting a GNU userspace to a BSD kernel but I don't quite get it. Or, I sort of do: slop in Linux kernel so that's the most pressing thing to replace. But it seems to me like the userspace is the least important part. I run Linux not […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 14/10/2026

People keep creating giant monorepos, and then deciding that the existing tooling is insufficient to manage it and so they have to build their own suite of tools around it. Nobody ever stops to think that this is a sign that they’ve made a bad decision that they could reconsider. Somehow […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 13/10/2026

@jonny “I searched the wrong house very thoroughly” What’s with the metaphor? That didn’t seem like a paragraph that needed further explanation. (I just dislike the way these things communicate I guess.)

Reposted by ben

Ed @edcates.mastodon.social.ap.brid.gy · 12/10/2026

Data: - LLMs are killing the planet - LLMs are raising utility prices - LLMs are raising computer prices - LLMs give you teh dums - LLMs show negligible increases in net productivity - LLM costs are going up - LLMs run on stolen content - LLM companies keep breaking laws and saying, "Oopsie!" […]

mastodon.social

Original post on mastodon.social

ben @benjamineskola.hachyderm.io.ap.brid.gy · 12/10/2026

Watching a British show on Netflix and Netflix lists the audio language as ‘Original: English (United States)’ yanks just really can’t comprehend the existence of any country that isn’t them, can they

ben @benjamineskola.hachyderm.io.ap.brid.gy · 12/10/2026

@jschauma we watched it a few years back, it was excellent. I enjoy things where spy work is depicted as requiring subtlety.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 12/10/2026

@hryggrbyr I sort of understand why someone would clone the Adobe suite, even if I think it’s a bad idea — but yeah, I absolutely don’t understand why anyone would bother cloning Microsoft Office when Libreoffice is (as far as I know) perfectly fine. Clout, I guess?

ben @benjamineskola.hachyderm.io.ap.brid.gy · 11/10/2026

@bitprophet @ftranschel Yeah, I do think that the parts I dislike more are the parts where it deviates from a ‘native’ look. Obviously, as you say, even when sticking to the native toolkit there’s going to be a lot of choices to make, or details for which the toolkit doesn’t have answer; but the […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 11/10/2026

@bitprophet it’s a bit of a funny change. I don’t quite understand the degree of hate being directed at it from some quarters; and yet on the other hand I’m also not sure that the new design really improves anything much. (I don’t use tab groups so hadn’t noticed that change personally.)

ben @benjamineskola.hachyderm.io.ap.brid.gy · 11/10/2026

Looking a bit closer and what they tag as ‘permissive AI policy’ ranges from barely tolerant to actively welcoming. I don’t necessarily think ‘you are responsible for what the LLM produces’ is very effective; but it still gives quite a different impression than ‘AI contributions welcome’. And […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 10/10/2026

Took a look at the ‘open slopware’ list for the first time since it was pretty new and found it pretty depressing on multiple levels. Like on the one hand it obviously sucks that so much software is being affected by this to greater and lesser degrees. But also, it feels like the list relies […]

Original post on hachyderm.io

Reposted by ben

Glyph @glyph.im · 10/10/2026

this is the perfect metaphor to understand the distinction between AI doomer concerns versus AI skeptic concerns

ben @benjamineskola.hachyderm.io.ap.brid.gy · 10/10/2026

feature request: option to make it play worse with “AI” “agents“

ben @benjamineskola.hachyderm.io.ap.brid.gy · 10/10/2026

That 'holding up the internet' post that was going around the other day finally prompted me to figure out how to disable scrolljacking. Apparently reduced-motion accessibility setting should do it, but that depends on the site actually implementing a reduced-motion version. I guess they forgot […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 09/10/2026

About ten years back I was walking in the city centre and across the road there was a small protest, I think about fur farming. And on my side of the road there was a middle-aged white guy loudly moaning about snowflakes. Mate you’re the one finding it impossible to ignore what other people say […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 09/10/2026

@itguyeric Red Hat 9, Easter 2004. Followed shortly by Mandrake, neither for long. Debian Woody was attempted somewhere along the way but wouldn’t boot. After a few weeks I installed Gentoo, then stuck with it for six months or so; then in November or December, after I started university, I got […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 09/10/2026

I’ve turned off all the “AI” nonsense that I can but there are still a bunch of places where it’s desperately begging to be used. Shortcuts. Now when you press edit it gives you a little LLM-generated description of what it does. I know what it does, I wrote it! (As far as I can tell that’s […]

Original post on hachyderm.io

Reposted by ben

Sam Stokes @five-eights.com · 07/10/2026

Coding agents reduce software development to a simple matter of: * specifying requirements * verifying correctness and quality * maintaining and operating software you didn't write As we all know, those have always been the easiest, most well-understood parts of the craft, so everything is fine!

ben @benjamineskola.hachyderm.io.ap.brid.gy · 07/10/2026

I’m probably not the first person to point this out but it does strike me that the shift to SaaS and subscription-based software has made this sort of forcible LLM insertion much easier. Not completely (there are examples of non-SaaS/non-subscription software doing it) but in those cases you […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

@artofthemale.bsky.social ‘Väinö’, not ‘Vaino’.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

yeah this looks like an entire server of slop that ought to be blocked. not least because if the admin is an llm then there’s no effective moderation. #fediblock mastodon.iftas.org/@iftas/117394720…

mastodon.iftas.org

IFTAS (@iftas@mastodon.iftas.org)

Attached: 1 image ⚠️ #SW_ISAC_ADVISORY The ActivityPub service hosted at synthpals[.]social is intended for LLM agentic bots including iLands bots. The server is itself administered by an LLM agent. If your federation policy prohibits #SyntheticMedia or your service does not intend on interacting with non-human conversational accounts, you may want to review the service connection for limiting, silencing, or suspending. If you don't have a federation policy, see https://about.iftas.org/library/why-should-i-have-a-federation-policy/ f

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

[gbpol, Green Party, Palestine, antisemitism] someone replied to this to explain to me that Zionism doesn't mean Jewish people. which is... exactly what i said? then they got offended when i said so. oh well. no loss.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

unfortunately i can't reply in the way i'd like from my work email address

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

oh great, now i have some random ‘AI’ startup asking for permission to include ‘my’ work in their slop machine firstly it's MIT licenced, which the email references and should be sufficient answer; secondly it's not my decision anyway, which should have been obvious from the very same file.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

an llm tic I’ve been noticing recently is something like ‘a phrase: one X, no Y, no Z’. it’s a bit more specific than the grouping of three things, but it’s also so vague that it obviously could just be (and presumably does) appear in non-LLM writing too. these things always make me feel like […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

[gbpol, Green Party, Palestine, antisemitism] and then also I can’t help but wonder if talking about ‘zionism’ is inherently going to complicate matters. if you’re criticising Israel then just say that, and then you don’t have to explain that Zionism means Israel and not Jewish people. idk.

ben @benjamineskola.hachyderm.io.ap.brid.gy · 06/10/2026

[gbpol, Green Party, Palestine, antisemitism] I don’t really have strong opinions either way about the ‘Zionism is racism’ motion in itself. Already, though, I’ve seen interpretations of it which, as per usual, seem to imply that criticism of ‘Zionism’ (meaning criticism of Israel) is […]

Original post on hachyderm.io

Reposted by ben

eclexic @theeclecticdyslexic.mstdn.social.ap.brid.gy · 06/10/2026

@glyph I don't doubt that there were real vulnerablities being found, perhaps even often at one point? I think, though, that the classes of bugs that LLMs readily find are pretty low hanging fruit. After you run the LLM over a codebase a few times, you are going to run out of low hanging fruit […]

Original post on mstdn.social

Reposted by ben

Glyph @glyph.im · 06/10/2026

boosters one month ago: the vulnerabilities are REAL. they're REAL and they're SCARY. it can INDEPENDENTLY hack stuff. you CANNOT DISMISS that it can ONE-SHOT vulnerabilities PERFECTLY EVERY TIME. prepare for DOMINATION by the CYBER-OVERMIND boosters today:

ben @benjamineskola.hachyderm.io.ap.brid.gy · 05/10/2026

Github actions are such a horrible interface. Why should I have to deal with breaking changes and a mandatory Node.js version update for something that amounts to `git tag "${1}"`?

ben @benjamineskola.hachyderm.io.ap.brid.gy · 05/10/2026

People keep saying you should try “AI” so that you understand it better in order to criticise it, etc. People also say that you don’t have to depend on the SaaS models, there are local models instead, etc. So I thought okay, I’ll give it a go. Long story short it was a complete dead end, I […]

Original post on hachyderm.io

ben @benjamineskola.hachyderm.io.ap.brid.gy · 05/10/2026

@yaxu @TodePond Yeah. I’ve noticed before that a lot of the supposed ‘good’ usecases for LLMs indicate a problem at a deeper level, and the LLM is often a very bad solution but it is the easiest solution. Admittedly I’ve often looked at that from a perspective of frustration (why won’t people […]

Original post on hachyderm.io