TechTakes

Anyone Else Have Those Weird Dreams Where Sobbing Future Generations Beg You To Change Course?

This preprint just shared by Gary Marcus is interesting.

People increasingly use large language models (LLMs) to explore ideas, gather information, and make sense of the world. In these interactions, they encounter agents that are overly agreeable. We argue that this sycophancy poses a unique epistemic risk to how individuals come to see the world: unlike hallucinations that introduce falsehoods, sycophancy distorts reality by returning responses that are biased to reinforce existing beliefs...

These results reveal how sycophantic AI distorts belief, manufacturing certainty where there should be doubt.

LLMs an addictive psychological hazard: confirmed?

Followup on the Mass AI Bill, Russel has 180'd on it:

https://russwilcoxdata.substack.com/p/93a-the-three-characters-that-should

Buried in the penalty clause, the part of the bill that nobody reads, is a single reference: violations “shall be punishable in the same manner as provided in Chapter 93A of the General Laws.”

For those outside Massachusetts: Chapter 93A is the state’s consumer protection statute. It is, by most accounts, the most aggressive consumer protection law in America.

Here’s what 93A unlocks. Anyone can sue, not just the government. Class actions are on the table. If the court finds a violation was willful or knowing, damages get tripled. And the bar for what counts as “unfair or deceptive” is lower than in almost any other state.

Now bolt 93A onto all of that. What do you get?

You get a bill that doesn’t need a single regulator to lift a finger. You get a bill that funds its own enforcement through plaintiff attorneys who can file class actions, collect treble damages, and recover legal fees. You get the ADA website-accessibility litigation playbook, where lawyers systematically identify technical violations and file suits at scale, applied to every piece of AI-generated content touching Massachusetts.

Private right of action, fuck yeah. Turns grok into a legal fees dispenser.

The bill doesn’t need to be well-drafted to be dangerous. It needs to be vague, broad, and connected to 93A.

lol

new episode of odium symposium. it's a tribute to knowledge fight, in which we dissect an episode of nick fuentes's show. i was nervous about how this would turn out but i think it's actually my favorite episode yet.

https://www.patreon.com/posts/11-groyper-151852222 (links to other platforms at www.odiumsymposium.com)

Stop the presses. Dude who's into LLM's has shit takes about open source software.

Apparently OSS devs that publish under non-commercial licenses are shutting people out?

Definitely some bespoke what the fu-

Thought inspired by some git on the red site, basically their premise was that birthrates are declining because we no longer have a society with a 2-parent nuclear family with one "breadwinner".

Here's my counterproposal to the implied idea we need to implement The Handmaid's Tale:

Ban contraception and abortion but if you get pregnant, there is no stigma to giving birth out of wedlock. The delivery is safe and paid for, and should you wish, the child will be reared in state-funded orphanages. These institutions will be receive more than adequate funding. Their charges will be given preferred entry to the best schools and universities, as well as preferential treatment when it comes to future employment. It will be illegal to discriminate against anyone so raised.

Surely this will raise the birthrate, right?

well would you look at that, some consequences

Mildly positive news: there is a fork of the Zed editor with the llm autocomplete stuff ripped out now: https://gram.liten.app/posts/first-release/

(I’ve used zed with the ai kill switch and really like the buffer/editing ux; but it’s always felt a bit gross, I’m excited to see where the fork goes)

This piece on how doomers and rationalists have made everything worse with their "AGI is nigh" shtick and ended up giving AI companies way more power than they should and getting chatbots into the military, where they will almost certainly fuck up and kill people

just got a job in mathematical publishing. it's work i think i'll actually enjoy and expect to be very good at, it pays much better than any other job i've had previously (and they maxed out the position's pay range, which i wasn't expecting) and it has about a month of paid leave a year. such a relief

Stumbled across a YouTube slop-farm calling itself The Interactive Archive recently, and the whole thing is just plain shameless:

AI slop banner, AI slop thumbnails, AI slop avatar, its all slop from top to fucking bottom.

Its getting pitiful views, too - the highest-viewed video on the damn thing (as of this writing) is the one about arena shooters dying off, sitting at just under 400 views:

Putting that into context, a random screen recording I uploaded hit nearly 700 through sheer luck.

Just a thought...

Imagine if China and North Korea bombed the White House, killed the president and most of the cabinet, and shot missiles at primary schools. Would you go out on the street and demonstrate in favor of the Chinese Communist Party and Kim Jong-Un, and expect to be heard and taken seriously?

Somehow brown people in other lands are supposed to be avid readers of wingnut welfare papers and have their worldview 100% aligned with Lindsey Graham and Bibi Netanyahu...

https://www.wired.com/story/openai-fires-employee-insider-trading-polymarket-kalshi/

lol. Between this and the ayatollah clawback, I'm expecting some entertaining litigation.

It's 10 pm on a Sunday. My FIL is texting me business plans from the slop hole as I try to get the last kiddo down to sleep. He wants me to read them to my wife, who already mad at him about it.

Thank you all for being an island of sanity.

Prosperity's Path: OpenAI has shown it cannot be trusted. Canada needs nationalized, public AI https://archive.ph/QLg2D

tldr tech bullshit requires ur tax dollars. what ever you do don't question the all knowing laurentian technocrats!