iiinova — where your digital self keeps evolving

virski.bsky.social @virski.bsky.social · 16h

There seems to be "I'm not a transhumanist" EAs who come up with... The same machine god type arguments, and who have close proximity to EAs.

100

virski.bsky.social @virski.bsky.social · 09/10/2026

Someone said that transhumanism means everyone would die. That's not true. There has been an idea or two from commenters in Silicon Valley that aren't entirely thought out, but then, I have seen ideas that are more thought out. Where someone might be getting stuck is that specific ideas require

100

virski.bsky.social @virski.bsky.social · 09/10/2026

Casey says he had a "meeting in Berkeley". A place that is notorious for AI doomsday cultists. Plus, he is known to be an Anthropic shill (and has a fiance who works there). #AI

000

virski.bsky.social @virski.bsky.social · 07/10/2026

Truth be told, this isn't the only area where I have seen someone saying that everything would happen at once. #AI

000

virski.bsky.social @virski.bsky.social · 07/10/2026

I've said this a number of times, but while a Singularity might work as a plot element in a fictional story, it is very unrealistic. #AI

000

virski.bsky.social @virski.bsky.social · 07/10/2026

I think TESCREAL is going too far. It implies everyone in Silicon Valley is sitting there gleefully thinking about their imminently becoming a cyborg or a mind upload. I've also seen transhumanists who don't think a so-called superintelligence is imminent.

131

virski.bsky.social @virski.bsky.social · 07/10/2026

The AI Doomsday Cult is mad that people don't buy their pronouncements about the imminent arrival of a machine god.

010

virski.bsky.social @virski.bsky.social · 07/10/2026

A medical doctor with a bit of psychology knowledge (who has historically made fun of disabilities) is upset that someone wasn't cordial about his membership in an AI doomsday cult. EA. #AI

000

virski.bsky.social @virski.bsky.social · 07/10/2026

I'm reading a take from someone supposedly at OpenAI and it's bizarre control freak colonialism cooked up by a cultist. #AI

000

virski.bsky.social @virski.bsky.social · 04/10/2026

I don't think people questioned AI Doom earlier because: 1. We didn't think there was an AI doomsday cult. 2. It seemed like an abstract thought experiment.

000

virski.bsky.social @virski.bsky.social · 02/10/2026

Google launched a prototype satellite into space. It's similar to the one with the GPU from earlier this year from Starcloud.

000

Reposted by @virski.bsky.social

virski.bsky.social @virski.bsky.social · 22/08/2026

Dramatic AI scenarios often assume people being incredibly stupid. What if we flip a coin to decide whether to fire the missile. And it assumes that after the missile misfires, that people will keep doing the same thing. #AI

001

virski.bsky.social @virski.bsky.social · 02/10/2026

If someone is concerned about this, then maybe they should question how it's being used, rather than talking about "AI Doom" which makes it sound like "the AI" is doing something (or that it's mystical). As mentioned previously, there are a number of reasons why this wouldn't happen. #AI

000

virski.bsky.social @virski.bsky.social · 01/10/2026

For a few years now, there has been talk of synthetic data being used to train models.

000

virski.bsky.social @virski.bsky.social · 01/10/2026

Someone's talking about overfitting? Overfitting is when a model fails to sufficiently generalize from the training data. It winds up being useless in practice as it can't handle cases outside of that. I'm talking about generalizing in a class, such as doing x.

100

virski.bsky.social @virski.bsky.social · 01/10/2026

Stanford was an early attempt at regulatory capture by Big Tech to try to undermine competitors. They followed in the spirit of former Facebook COO Sheryl Sandberg, who used to work in politics.

100

virski.bsky.social @virski.bsky.social · 01/10/2026

It's actually worse than them saying by 1000%. They were saying that they were "more of a moderate" and that they spoke to someone who thought the GDP would grow by ten thousand times per year. Both numbers are dubious.

000

virski.bsky.social @virski.bsky.social · 01/10/2026

Another issue with Stanford is that there are times when someone is like "algorithmic censorship is fine" when arguing against a kind of privacy violation. It's like they're talking from the perspective of a Big Tech company.

000

virski.bsky.social @virski.bsky.social · 01/10/2026

While there have been times when Stanford has been critical of government censorship, I am concerned about essentially PR rationalizations for policies making their way to Stanford (when quite a number of these policies are bad and censorious).

100

virski.bsky.social @virski.bsky.social · 01/10/2026

Stanford has close ties to Silicon Valley. From an economist making dubious claims that the GDP would grow by 1000% per year due to #AI to apparent attempts to trip competitors. When the former CEO of Google DeepMind (he was moved to a

100

virski.bsky.social @virski.bsky.social · 01/10/2026

There is a sense that this is more of a "we are going to take steps to improve our security" event but there was an open-ended aspect that I wanted to comment on.

000

virski.bsky.social @virski.bsky.social · 01/10/2026

Commenting on the SI Agreement. We should avoid associating "developing AI" with such and such measures. Something like that might be obsoleted at some point by a system (maybe, it'll be narrower in the areas that it covers) intelligent enough to do what the user wants.

100

virski.bsky.social @virski.bsky.social · 30/09/2026

Reminder that a number of media outlets were for anime, until Trump won the 2016 election, and they decided they hated the Internet. Doing that though only alienated people and made it more likely Trump would win again.

000

virski.bsky.social @virski.bsky.social · 30/09/2026

I'm not saying that people aren't fans of something, especially something popular. On the other hand, saying that everyone is fansexual or something esoteric. It seems to me that someone is looking for a box to put people into. This reminds me of a time a number of years

100

virski.bsky.social @virski.bsky.social · 30/09/2026

To be fair, Silicon Valley investing in immortality did lead to a treatment for glaucoma (which can lead to blindness) where there was none before.

000

virski.bsky.social @virski.bsky.social · 30/09/2026

Apparently, AI is being rebranded as SI.

000

virski.bsky.social @virski.bsky.social · 30/09/2026

At one point, there was a word that means "behaviors", especially during that time, and in the context of crimes. A Canadian guy ascribed meaning to it that it didn't have.

100

virski.bsky.social @virski.bsky.social · 30/09/2026

Another successful launch of SpaceX's Starship vehicle. I'm hearing that the previous one was sub-orbital and this one was orbital. The landing of the rocket was smoother this time round.

000

virski.bsky.social @virski.bsky.social · 30/09/2026

In terms of transhumanism, dealing with such crimes would be doable. Suppose someone doesn't process an event that happens to them, or they lower the weight of it. Also, there would be wiser decision making. Wise is a vague concept, I know.

100

virski.bsky.social @virski.bsky.social · 30/09/2026

Someone can't really say that a vast group is all one particular kind of person. The world is not so simple.

100

virski.bsky.social @virski.bsky.social · 30/09/2026

Humans like to dance around apparent taboos. For humans, a story that someone is something else sounds good. They won't have to talk about it. And it is true that someone wouldn't want to go commit crimes. However, notice how someone reads out a specific thing that someone didn't say.

100

virski.bsky.social @virski.bsky.social · 30/09/2026

Someone said that Dr. Saito said that very few who are interested in "lolicon" find minors attractive, lower than the general population. This is a misquotation. Dr. Saito did not say that. What he said is that there are *very few sex crimes*, lower than the general population.

000

virski.bsky.social @virski.bsky.social · 27/09/2026

There were a few people conflating accessing an unprotected page with hacking. If someone comes across a page that someone hasn't protected, that's not hacking. A bot might access a number of such pages as it handles a site. #AI

100

virski.bsky.social @virski.bsky.social · 23/09/2026

Anthropic has released a new model. Notice how Anthropic engaged in doomsday rhetoric prior to doing so. They don't like that OpenAI released a new model. They dislike competition.

001

virski.bsky.social @virski.bsky.social · 22/09/2026

They likely went over thousands or tens of thousands of problems before finding one with an ambiguous framing (that has long been controversial) to try to "solve".

000

virski.bsky.social @virski.bsky.social · 22/09/2026

Mathematicians posted a proof debunking OpenAI's new model solving Navier-Stokes. Apparently, OpenAI took an unnatural approach which looked for ambiguity in the framing of the question, rather than coming up with something actually useful.

100

virski.bsky.social @virski.bsky.social · 19/09/2026

Anthropic seems to want to monopolize bio. Even if we ignore that they tend to hype up their tech, and they have no expertise in bio, they would end up going up against Big Pharma. Even if some AI was great, Big Pharma could just train their own. #AI

101

virski.bsky.social @virski.bsky.social · 18/09/2026

Someone said that if someone uses an unreliable LLM for air traffic control, then that would cause issues. However, doing so would be incredibly stupid (and this example is very contrived). #AI

000

virski.bsky.social @virski.bsky.social · 17/09/2026

Microsoft's so-called MAI "Code of Conduct" is really just a marketing piece. It talks of content that Microsoft doesn't want to handle as if Microsoft is doing something "ethical", rather than trying to market a rather dry and generic policy for a piece of productivity software. #AI

110

virski.bsky.social @virski.bsky.social · 14/09/2026

I'm hearing that they have a weird background. That they're funded by AI doom, that they're in the same orbit as AI doom, that they anthropomorphize. #AI

000

virski.bsky.social @virski.bsky.social · 14/09/2026

Anthropic has suggested to get an org run by, uh, their investors to spy on their competitors (and to make statements about them). #AI

000

virski.bsky.social @virski.bsky.social · 12/09/2026

Someone compared a LLM to neurons. The brain is nothing like a LLM. In the brain, neurons are connected to a thousand other neurons (or more). That is a thousand inputs and one output. They communicate with another neuron via a spike (a surge in voltage of 40mV). #AI

100

virski.bsky.social @virski.bsky.social · 12/09/2026

Someone said something like "I don't understand anything, I'm not thinking of any scenario, I assume the AI will just magically destroy the world." There are also mentions of "I am idiot", which seems like projection, that because they think that, that they think everyone else is an idiot too. #AI

000

virski.bsky.social @virski.bsky.social · 12/09/2026

With "the grid", if someone is that concerned, the grid could adopt a practice called "air gapping" where the devices are disconnected from the Internet. Also, there have been calls to improve security here for over a decade. Let's not pretend this is new. #AI

000

virski.bsky.social @virski.bsky.social · 12/09/2026

There doesn't have to be a sophisticated AI for "the grid". There are already algorithms that are more reliable and which are about as good as it's going to get. #AI

000

virski.bsky.social @virski.bsky.social · 11/09/2026

As I've said before, one argument boils down to assuming that someone will do something incredibly stupid like deciding that a missile launch will be determined by a coin flip, then after it misfires, they will keep doing the same thing. #AI

100

Reposted by @virski.bsky.social

Mike Drucker @mikedrucker.bsky.social · 09/09/2026

You’ve gotta use AI because it’s the future. But also it’ll take your job. But also it’ll create a paradise on earth and solve most of our problems. But also it’ll end humanity soon. But also you’ll get left behind if you don’t use it right now.

12068181638

virski.bsky.social @virski.bsky.social · 10/09/2026

While I do think they buy into their own hype, and Anthropic's ties to an apparent doomsday cult isn't something that inspires confidence, there is something strange about it. Consider that people who work at companies like this might own stock in them.

210