Except the adverse outcomes are entirely predictable. Not the exact nature of particular exploits, but I like the analogy Cal Newport keeps using in his videos: if you strap a weed whacker to a dog, you shouldn't be surprised if it then jumps the fence and runs around cutting people's ankles and other random bad stuff.
Plenty of people on that list maintain serious production systems or open source projects in use by a wide variety of people. It's not just some nerds tinkering in their basement.
> Even if they're producing code by hand, surely they have LLM helpers by now.
Nah. Most of my coworkers got so annoyed by VS Code injecting copilot and other "helpers" that they have switched to Kate or Vim.
There are still a lot of places that value engineering excellence, and plenty of people who value doing the actual work and knowing that it is correct, instead of delegating to an LLM and hoping.
You are living in a bubble. I've never used LLM helpers, and in my circle nobody does willingly. Those who are forced by their bosses to use such tools really enjoy in their free time coding by hand without a stupid robot watching over their shoulders.
We are billions of people opposed to AI as an ecocidal tech and a heteronomous political project, including much of the IT sector. Just because you only follow grifters and scammers trying to profit from the AI bubble doesn't mean "surely" we all follow those trends.
I never pretended otherwise, unlike you. There are multitudes of humans and cultures, but you are reducing us all to your overhyped narrow silicon-valley view of humanity. Just because a few million people adhere to your AI death cult doesn't mean the rest of us "surely" do.
I keep feeling "surprised" that people who seemingly would know better post stuff that obviously sounds like an LLM. But then I remember, you and I aren't the audience. The audience is investors. And investors seem to love LLM-voice, because…in large part…the LLMs sound as vapid and unhelpful as the VCs.
I and a significant number of people I know and interact with on a regular basis are creating and publishing stuff for the human indie open web much like we have for years and decades. The underlying specs and hardware and tools and such get better (sometimes, the ones which haven't enshittified), but the techniques largely stay the same.
Huzzah! This is excellent news. Safe spaces and healthy communities are borne from enforcing strict moderation to keep the bad actors out. This is an important step forward for the platform.
... and leave it completely up to their own personal judgement on when and how to apply their rules, and on public forums people who asked for more clarity on "boundaries" were called "creeps" and other insults.
Fantastic! Codeberg just keeps getting better and better, properly moderating a healthy community which does not welcome the technolibertarian dystopia we've all been suffering through for far too many years now.
I'm not surprised many here are upset. Many here don't seem to understand the outrageous harms that a totally out-of-control tech industry has been foisting upon the planet. It's time to take a stand for a better path forward.
actual open source crypto projects are the opposite of a techno dystopia. ethereum wants to build a global root of trust (like iana but decentralized), railgun and monero are doing anonymous payments, and all of them let you avoid taxes. which is good if you live in a country where your taxes go to foreign genocide and the war on drugs.
all the harms come from for-profit companies that use crypto as a cover, like polymarket or stake. the contracts are on chain and might be open licensed but one company controls everything in practice with admin rights. and they make it hard to access directly without going through the proprietary frontend hosted on their servers.
the only ones who would ever want to host their code on codeberg are real community projects, because the corpos are all happy with github.
Only people actually do care about the process by which the products they use were created. It's our job, as experts who build things, to educate our audiences why and how we build what we build.
Meaningful percentages of people already hate the obvious slop being thrown in their faces. The more they learn of the slop infecting the inside of the digital products they use—and the resulting enshittification—they will seek out worthy alternatives.
TLDR I care more about the outcome not being slop than I do an LLM being involved in its creation.
Sure but not everyone is as process driven, some are more outcome driven. Personally I care more about the outcome of your software (its quality) than I do the process you used to create it (LLM gen? transpired from some other language?).
Back in the day when I was into reversing MacOS apps (during the PowerPC era) there were these cross platform frameworks that were used sometimes that I really loathed. I did not like them because
- Reversing: The binaries were a spaghetti garbage mess that (unintentionally) amounted to some serious obfuscation
- User Experience: Ugly (no native elements), Slow (because of the aforementioned spaghetti), and Bloated (took up a lot of space back when that came at a premium)
The thing is though, I would not care if they had made these in Xcode Obj-C natively and just did a poor job of it. To me the end result was indigestible, and the process by which it was derived was neither here nor there. Even without LLMs, my world at that time was abound in generated slop that I did not like.
I think it's totally acceptable to be process driven. Much interest in art is based around the process or lived experience of the artist more than the art itself. I think it's mostly an aesthetic or cultural difference which side of the spectrum (if any) between process and outcome you live on.
reply