Follow up to the article from this morning.

[tomjennings@tldr.nettime.org]

@tante @pluralistic

As soon as google first used their "don't be evil" phrase I called bullshit. Whats "be"? Don't *do* evil -- now that's at least possibly accountable.

I don't buy from Amazon. It's a shit company with exploitive practices, done and done. Purity? I don't even care. It harms me, a dollar at a time.

I don't want to *think* using bad tools. LLMs are inherently fascist technology, the roots of them go back to post-ww2 mind/body split christian ideologies, force-fit for six decades where they don't fit. Of course other systems based on Shannon's statistical views of (english) language could have gone other ways; and still could go other ways; just as Amazon could become a good business in the future (I suspect the latter is less likely, even, than the former).

If Cory uses a small worm in an underground pool to do grammar and spell check, that's a great use of the underlying ideas in a large language model system. I wouldn't, but right here in this sentence I cannot tell you why, other than, I just don't trust any of the people involved in the makings. We're all living in dangerous times, near many sharp and dangerous objects.

[reflex@retrogaming.social]

@tante @pluralistic When people point out to him how or why something is a bad idea or a disservice to his readers he typically either argues with the point on a technicality while avoiding the actual problem, or he ignores them entirely.

Maybe you just noticed it due to how he disingenuously framed his LLM usage, but for many of us it's the latest in a line of examples and his dishonest defenses of harmful tech.

[vksxypants@mastodon.social]

@tante this is the first time I've read your blog, and now I've read two in 30 minutes. I can't stress enough how wonderful both posts were -- incredible stuff.

[tante@tldr.nettime.org]

@vksxypants don't get used to it, I am as much hit and miss as everyone else ;)
but thanks

[taaldader@waag.social]

@tante
Favorited the first article. Don’t think I “needed” the second one, but nonetheless favoriting it also, because I appreciate the move to preempt a possible degradation of the quality of discussion, as well as the focusing on the fights ahead.

[pythonbynight@hachyderm.io]

@tante @pluralistic

> And if we demand perfect, flawless ideological congruence we will lose.

^ that tends to be a guiding theme for much of my writing. Thanks for the follow up!

[alatiera@mastodon.social]

@tante @pluralistic Great post. The whole thing about consumption purity is incredibly annoying and misguided.

The thing I rarely see pointed out, is how this usually stems from a place of powerlessness and complete alienation from every other aspect in life. The people demanding this are the ones who's identity is now defined entirely by their consumption habits and they are so lost that this is the only way they can imagine demanding any change, hoping to reclaim some autonomy in their life

[tomjennings@tldr.nettime.org]

@tante @pluralistic

Bigot Shockley's beliefs -- who if you recall was team manager that did little of the actual work yet took the credit -- those beliefs aren't embedded in the materials science that made a practical transistor. And Shockley ended up being a footnote in the industry, with his betting the house on four-layer diodes and other forgotten things. And in the ~70 years since, all of those earliest players have been subsumed into a nearly world-wide industry of a scale that dwarfs any imaginings of 1950.

The poisonous tree of so-called AI and LLMs is recent, fresh, much more intentional, guided, and virulent. The corporate proponents have demonstrable ill intent. They do harm, and seem to be inherently financial scams. None of this is an incidental side effect of bigoted personalities; the taint is built-in. Built into the methodologies of development and deployment, if not the design and goals. Personally I find it very difficult to isolate it from the work itself.

I'm not condemning Cory, but I find the arguments in favor of tolerating the associations with the code's developers difficult to accept. I'm wrong often.

But I've been fairly allergic to capitalist technologies for a long time, dropped out decades ago, so filter my words through that.

[dgodon@mastodon.online]

@alatiera @tante @pluralistic agree it’s a great article and critique. The discussion about consumption purity was a good reminder to myself

[seachaint@masto.hackers.town]

@tomjennings @tante @pluralistic If we're comparing diodes to LLMs then the embedded politics is just so different. I can rely on a diode doing the same job for me as for a nazi. The same is not true of LLMs.

And that's true of any two use-cases, and I don't know if we even have an academic framework yet to figure out how to fix that "problem" (supposing for a moment that the billionaires even want it fixed).

Fundamentally: this tech is unnecessary, and it's boiling the planet. That's enough.

[rpbook@gts.phillipsuk.org]

@reflex @tante @pluralistic what's wrong with Kagi? Most people seem to love it.

[malteengeler@legal.social]

@tante As I read @pluralistic, his goal is not to overcome capitalism but to reform it into a more "ethical, good version of capitalism". If one, however, actually fights to overcome capitalism there are significant differences between these struggles. To recognize that does not mean he wont' be an ally on the way but it also does not require calling him more than he is: A very smart writer who entertainingly hates of some symptoms of capitalism without questioning it fundamentally.

[pluralistic@mamot.fr]

@malteengeler @tante that is a complete misreading of my work. I am by no means interested in preserving capitalism. I think markets can be useful allocation tools but as I have said multiple times in many ways markets are not the best or only way to do allocation.

[sorenhave@mastodon.nu]

@tante I didn't read neither the post by @pluralistic nor your criticism of it, but I fully agree with your point that we need to be able to disagree now and then, as long as we agree on the overall issues.

@BrentToderian also put it well here: https://fed.brid.gy/r/https://bsky.app/profile/did:plc:ny5w2k6brb2kdubxnevke6li/post/3mezntfzlb22p