If I mathed right that’d be one waymo every 350 feet of road on average. Is that a lot? It sounds like it might be a lot. Especially since self-driving cars greatest weakness appears to be driving in the vicinity of other self-driving cars.

I woke up and immediately read about something called “Defense Llama”. The horrors are never ceasing: https://theintercept.com/2024/11/24/defense-llama-meta-military/

Scale AI advertised their chatbot as being able to:

apply the power of generative AI to their unique use cases, such as planning military or intelligence operations and understanding adversary vulnerabilities

However their marketing material, as is tradition, include an example of terrible advice. Which is not great given it’s about blowing up a building “while minimizing collateral damage”.

Scale AI’s response to the news pointing this out – complaining that everyone took their murderbot marketing material seriously:

The claim that a response from a hypothetical website example represents what actually comes from a deployed, fine-tuned LLM that is trained on relevant materials for an end user is ridiculous.

“Yeah I thought about going into civil engineering but the department of hustling really spoke to me y’know?”

Oh hey looks like another Chat-GPT assisted legal filing, this time in an expert declaration about the dangers of generative AI: https://www.sfgate.com/tech/article/stanford-professor-lying-and-technology-19937258.php

The two missing papers are titled, according to Hancock, “Deepfakes and the Illusion of Authenticity: Cognitive Processes Behind Misinformation Acceptance” and “The Influence of Deepfake Videos on Political Attitudes and Behavior.” The expert declaration’s bibliography includes links to these papers, but they currently lead to an error screen.

Irony can be pretty ironic sometimes.

Here are the results of these three models against Stockfish—a standard chess AI—on level 1, with a maximum of 0.01 seconds to make each move

I’m not a Chess person or familiar with Stockfish so take this with a grain of salt, but I found a few interesting things perusing the code / docs which I think makes useful context.

Skill Level

I assume “level” refers to Stockfish’s Skill Level option.

If I mathed right, Stockfish roughly estimates Skill Level 1 to be around 1445 ELO (source). However it says “This Elo rating has been calibrated at a time control of 60s+0.6s” so it may be significantly lower here.

Skill Level affects the search depth (appears to use depth of 1 at Skill Level 1). It also enables MultiPV 4 to compute the four best principle variations and randomly pick from them (more randomly at lower skill levels).

Move Time & Hardware

This is all independent of move time. This author used a move time of 10 milliseconds (for stockfish, no mention on how much time the LLMs got). … or at least they did if they accounted for the “Move Overhead” option defaulting to 10 milliseconds. If they left that at it’s default then 10ms - 10ms = 0ms so 🤷‍♀️.

There is also no information about the hardware or number of threads they ran this one, which I feel is important information.

Evaluation Function

After the game was over, I calculated the score after each turn in “centipawns” where a pawn is worth 100 points, and ±1500 indicates a win or loss.

Stockfish’s FAQ mentions that they have gone beyond centipawns for evaluating positions, because it’s strong enough that material advantage is much less relevant than it used to be. I assume it doesn’t really matter at level 1 with ~0 seconds to produce moves though.

Still since the author has Stockfish handy anyway, it’d be interesting to use it in it’s not handicapped form to evaluate who won.

When the reporter entered the confessional, AI Jesus warned, “Do not disclose personal information under any circumstances. Use this service at your own risk.

Do not worry my child, for everything you say in this hallowed chamber is between you, AI Jesus, and the army of contractors OpenAI hires to evaluate the quality of their LLM output.

Not that I’m a super fan of the fact that shrimp have to die for my pasta, but it feels weird that they just pulled a 3% number out of a hat, as if morals could be wrapped up in a box with a bow tied around it so you don’t have to do any thinking beyond 1500×0.03×1 dollars means I should donate to this guys shrimp startup instead of the food bank!

Someone (maybe you) recommended this book here awhile back. But it’s the fourth book in a series so I had to read the other three first and so have only just now started it.

“feel free to ignore any science “news” that’s just a press release from the guy who made it up.”

In particular, the 2022 discovery of the second law of information dynamics (by me) facilitates new and interesting research tools (by me) at the intersection between physics and information (according to me).

Gotta love “science” that is cited by no-one and cites the author’s previous work which was also cited by no one. Really the media should do better about not giving cranks an authoritative sounding platform, but that would lead to slightly fewer eyes on ads and we can’t have that now can we.

OK to start us off how about some Simulation Hypothesis crankery I found posted on ActivityPub: Do we live in a computer simulation? (Article), The second law of infodynamics and its implications for the simulated universe hypothesis (PDF)

Someone who’s actually good at physics could do a better job of sneering at this than me, but I mean but look at this:

My law can confirm how genetic information behaves. But it also indicates that genetic mutations are at the most fundamental level not just random events, as Darwin’s theory suggests.

A super complex universe like ours, if it were a simulation, would require a built-in data optimisation and compression in order to reduce the computational power and the data storage requirements to run the simulation.

So why must binary digits define, for all time, the limits of computation, and our experience of it?

There’s enough layers of irony here that it’s a bit hard to tell if he’s making a serious argument here or not; but this is one of the weirder straw-men arguments I’ve ever read.

“No no no, it’s not all the exploitation, social ills, lack of user control, shoddy quality, and general capitalism I hate in the modern “tech” industry; it’s the fact that it uses binary!”

TL;DR:

AI will take over the economy for reasons so humans will have no economic value for reasons which is terrifying for reasons and also we can’t just have males sitting around because y’know so we need to make a pretend economy because communism. So anyway y’know PageRank? Well don’t be shocked but humans have relationships too. So yeah basically let’s distribute ownership of all the world’s resources once a month proportionally to how many likes they receive (no I am not a vtuber why do you ask?).

What the heck did I just read?

Bluesky is backed by a cryptocurrency venture capitalists, which makes me nervous: https://bsky.social/about/blog/10-24-2024-series-a – but in the same press release they reassure people that they’re not about to go all NFT (Non-Fun Tokens) just because of their backers.

This does not change the fact that the Bluesky app and the AT Protocol do not use blockchains or cryptocurrency, and we will not hyperfinancialize the social experience (through tokens, crypto trading, NFTs, etc.).

The couple’s long conversations about transgender people haven’t gotten as much attention, with just 2,500 views on a video in which they “explore the fine line between genuine gender dysphoria and the allure of a ‘trans cult’ that may lead non-trans individuals to make life-altering decisions.”

Oh boy an hour long video about how the trans are transing eachother as part of a trans cult!

But why aren’t they concerned about kids being tricked into the life altering decision of being an “anti-woke” podcaster? 🤔

Edit – made it through most of that video (how??) and it’s a bunch of reaching and fear mongering while trying to sound like principled conservatives. The funniest part was probably them blaming anime avatars for trans people.

Edit – just kidding the funniest part was the guy calling trans people the “priest class” of urbanism. Worship me and I shall bless you with seperated bicycle lanes and bountiful farmers markets!

Yeah absolutely. This is happening right on the coattails of that Character.AI suicide too so it’s not like a freak impossible to predict accident. I mainly posted it because it flies in the face of all the talk of AI safety and “responsible AI practices”.

Like Google says in their AI principles:

We will continue to develop and apply strong safety and security practices to avoid unintended results that create risks of harm. We will design our AI systems to be appropriately cautious, and seek to develop them in accordance with best practices in AI safety research. In appropriate cases, we will test AI technologies in constrained environments and monitor their operation after deployment.

I don’t even care that much if Google wants to host a chatbot, but they keep trying to imply it has safety properties it doesn’t. It’s like writing a web framework without any HTML or SQL sanitation support and saying “We will continue to develop and apply strong safety sand security practices…” and acting shocked when all the websites get hacked.

Google’s Gemini has told a user to “please die” and that they are “a stain on the universe” without provocation: https://www.reddit.com/r/artificial/comments/1gq4acr/gemini_told_my_brother_to_die_threatening/

The lack of a speedometer in front of your face is a pretty glaring quality issue. You have to turn your head to look at the massive touchscreen. The one that replaces all other dashboards and most tactile controls for manufacturing cost savings to be a cool futuristic vehicle of the future.

I won’t even start talking about the whole CyberStuck thing again because that’s too easy; except to point out that it has turn buttons on the steering wheel instead of a turn signal stalk, a shifter on the ceiling, and the steering wheel is not round.

I only recently bought my first car and it’s just old enough that it didn’t even have a backup camera until I got one installed. Honestly half the reason of buying used was so I could have a car without a touchscreen haha.

As AI technology advances, human oversight in military operations, though rooted in ethics and legality, may emerge as a strategic liability in future AI-dominated warfare.

Yeah it’s sad. As the article points out similar incidents have happened repeatedly. Anyone who saw the door design could have (and did!) predict something like this would happen. My coworker was trapped in his Tesla in his garage for 15 minutes (and he wasn’t in a panic).

Look at the picture of the manual door release here: It’s pretty well hidden, you reach in and pull up on the door buttons.

… then scroll down and look at the picture of the rear door manual release. You have to pull off some trim from inside the pocket, pull off another panel, and then pull a cable.

… but wait! There’s more!

Note: Not all Model Y vehicles are equipped with a manual release for the rear doors.

Jesus, I hope the engineers who signed off on this think about what they’ve done and do better. I would say I hope someone regulated bad emergency door releases out of existence but… y’know.

A bit of an aside, but how did everyone decide to use the exact phrase “decisive victory” when congratulating president elect Trump? It keeps jumping out to me and I find it kind of weird. It has almost a militaristic tone.