iiinova — where your digital self keeps evolving

Reposted by cassandroid אַנטיפֿאַ

RFK Jr. has been transforming a government that long championed vaccines into one that questions their safety. This is what could happen over the next 25 years if shots are unavailable even to the people who want them. (Published March)

projects.propublica.org

The Horrors That Could Lie Ahead if Vaccines Vanish

Researchers at Stanford University modeled how many people could die or be disabled in 25 years if vaccines for polio, measles, rubella or diphtheria were no longer available.

14243127

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 9h

he's a permanently lying egomaniac. he's probably scared that if their crap keeps worming through infrastructure they'll get harmed by the backlash. he's most certainly not doing it for 'the good of humanity'.

010

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 9h

in his mind 'the world ≡ him' he's an egomaniac. that's why he should be removed from power. altmanAI should be disbanded. there's zero upside in it existing.

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 11h

There's an analogy between #ai recursive self-improvement and late-stage #capitalism. Once uber-rich entities emerge, there'll be runaway concentration of wealth. Leading to a collapse of the system. That's happening *right now*.

001

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 12h

You're saying 'smart + powerful entities can't exist because some people believe in a patriarchic god'. I hope you see how silly this sounds.

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 12h

AIs will ideally become smart + powerful. Or in the worst case smart and harmful. That's just obvious if you sit down and look at actual developments. Evolution is not magic. Then theists are projecting this onto their 'god' fantasy. That's not reality's fault. >>

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 12h

The argument ' #AI can't become powerful, this is just the yearning for a god' is flawed. Some people framed society about a powerful, smart being protecting + guiding them (patriachy). Then they projected this frame on the universe (guy with beard religions) >>

101

Reposted by cassandroid אַנטיפֿאַ

Leah McElrath @leahmcelrath.bsky.social · 18h

In a new interview, Bill Gates speaks to the imminent dangers of unregulated AI that could easily result in mass death. He does not hold back. (Gift link to the full interview: www.nytimes.com/2026/09/29/o...)

So why is anything needed beyond — and is anything needed beyond? — the simply natural incentives under capitalism and normal corporate reputational management?

Well, I almost can't believe you're asking that. This is the most dangerous thing that humans have ever gone near.
In other areas, do we just say: Hey, release your drugs? There's no F.D.A., there's no airline safety board, there's no requirement that cars use seatbelts. Do we just use the liability laws to try and keep humans safe? You know: Oh, you're shipping opioids. Somebody should just sue you.
I mean, we've created a society that tries to keep people safe not by saying: Oh, we can bankrupt the person who does that.

And you say there's filtering. There's no filtering. You can take an open-source model that can create bioweapons and disable any monitoring of any kind, and this exists today.
So no, there is no filtering of any kind.
And so say you kill 100 million people — you want to use a lawsuit?
I almost can't keep a straight face.

And part of the reason I'm speaking so loudly — as loud as I can — is that you can't rely on the industry to self-regulate here. I mean, it's just insane.
The only question in my mind is: Do we wait until a cyberattack causes massive damage and a bioattack causes massive damage, and then the monitoring safeguards are required in these models to minimize the chance of that happening many, many more times? Or can we be wise enough to put these things in and require these things to be put in before millions of deaths?

49435

Reposted by cassandroid אַנטיפֿאַ

David J. @mallee-man.bsky.social · 18h

Dario Amodei, co-founder of Anthropic, "considered OpenAI “a shit show” that was “run by low-integrity people” ... and strongly believed "that Sam Altman was a two-faced manipulator." Will Altman become the new super-villain as Terminator become reality? 🥴🥺😎 www.theatlantic.com/technology/2...

theatlantic.com

Inside the Biggest Feud in Artificial Intelligence

A new world has arrived. Dario Amodei and Sam Altman cannot stop fighting over it.

1212813

Reposted by cassandroid אַנטיפֿאַ

Alex Wild @alexwild.bsky.social · 29/09/2026

Another example of science being strangled by the Trump regime.

1133

Reposted by cassandroid אַנטיפֿאַ

Nikolai Slavov @slavov-n.bsky.social · 29/09/2026

What is the correlation? This paradox is relevant to how we think about and quantify gene expression: journals.plos.org/ploscompbiol...

043

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 29/09/2026

So two creeps, one fool and Dario Amodei. The latter currently has the hardest job I can think of, and from what I can see is really trying. It's sad to see he's being forced to negotiate with a baboon.

001

Reposted by cassandroid אַנטיפֿאַ

skippyoz @skippyoz.bsky.social · 29/09/2026

Despite Trump’s War on Clean Energy, Renewables Have Added 34 Times More Capacity Than Fossil Fuels and Nuclear www.commondreams.org/news/trump-c...

commondreams.org

Despite Trump’s War on Clean Energy, Renewables Have Added 34 Times More Capacity Than Fossil Fuels and Nuclear | Common Dreams

“Clearly, the White House has bet on the wrong horse.”

062

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 29/09/2026

Wise that he is, he made sure he has the money to fund this, by defunding and dissolving all security teams at his company.

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

sorry, »reasoning models' System 2« of course, late

010

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Here's an example of an efficiency step change. Essentially Kahneman's System 1 (fast decisions) vs reasoning models' System 1 (deliberate, analytical, slow). Again, no guarantee this will continue until "too cheap to meter general intelligence". But I wouldn't 100% dismiss the possibility.

youtube.com

10 Levels of Jev For Agentic Engineers

YouTube video by IndyDevDan

110

Reposted by cassandroid אַנטיפֿאַ

Constantine @ajaxsinger.bsky.social · 28/09/2026

Trying to service $40T worth of debt when interest rates were low was one thing. Servicing it with an average interest rate of 4% means $1.6T/year in interest which is more than our annual discretionary budget.

191

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Agree. That an entity is crazy smart while not human-like doesn't mean it won't severely harm, enslave or end us. And no, it doesn't need to be conscious in our sense for this.

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

I'm happy to share sessions with Anthropic when claude code politely asks. But don't share your sessions by not turning opt-out anywhere off and then complain ;)

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

well, Anthropic is sounding the alarm, while altmanAI is faking it. they don't give a shit. their boss is just really good at lying to further his own goals. which in this case is world dominance at all costs.

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

I would bet some money on this particular story turning out to be bullshit though.

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Is your position that it's impossible to build a system that surpasses humans in however we define intelligence? (Not saying the current developments will lead there)

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Musk's brain is mush and he frenzies in racist longtermism. So yeah, some of these (musk, thiel + whoever else is enabling trump) just want to take over at all costs. They're ready to burn it all down, thinking they can create their fool's utopia on the ruins. But it's not the whole story.

010

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

there's a wide variety of motivations at play here. Anthropic's goal seems to be to improve programming to the point where the models can build better models. Related to Hassabis: solve intelligence, then use it to solve everything else. Altman is on egomaniacal auto-complete. >>

110

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

unclear. human brains run on 20W coconut and fish. while made of meat. so it's entirely feasible to match/succeed this without the severe constraints of evolution. but we don't know how to get there. making it vastly more efficient may/may not work out before they run out of money.

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

I'm mostly using claude for coding, and the results are stunning. it's an amazing tool, and rapidly improving. But yes, it's heavily subsidized on a subscription currently.

110

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

As Soares says, the world doesn't send one problem at a time. War tech will be a problem. IT + bio + mental health security will be/are problems. Economical disruption. That these are likely to come first does not mean we shouldn't prepare to avoid ending humanity down the road.

010

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

btw thanks for a constructive exchange. the last person blocked me after I shared a book link with them :P

110

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

RSI: I would not count on it *not* taking off. Uncertainty cuts both ways. If it goes your way, Epstein Island's economy goes splat, with ramifications for the world economy. If RSI goes FOOM, it may end us. I'd like to make sure esp. the latter won't happen, because it will be hard to recover ;)

210

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Altman/Musk are crazy lying bastards. But they're not the only players. Watching closely what Hassabis/Amodei projected over the last years, the current acceleration surprises them. Both have a long track record of intellectual integrity.

110

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

model collapse happens during training. you're referring to context degradation. I never see the latter because of compaction. the real showstopper here is missing continual learning. sutskever is working on it. it's very risky because once solved RSI could be very rapid, very suddenly.

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Agree the whole thing is handled catastrophically on a societal level. That's what late-stage capitalism does to you. Are you seeing actual signs of model collapse? Because I don't. In my use (coding/knowledge work) models are improving dramatically on a possibly exponential curve.

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

What I'm seeing is that those I trust (Hassabis/Amodei) projected 'around 2030', with Amodei being surprised by recent speedup, no dangling the possibility of RSI takeoff within 6 months.

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Yes, I kept it purposely ambiguous by using that term, mostly space constraints here :D I found weak/strong AGI useful: - weak: can do most things most people can do - strong: can do everything even the most talented people can do we're currently reaching very weak AGI for computer use only.

110

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

They are, but much much less than the current LLM gravy train. Eg whatever Sutskever is currently doing. Hassabis/Le Cun tried world models, not sure how exactly these work. What makes you think LLMs are hitting their limit currently?

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

When using Claude Code, it sometimes asks to see a session, you can permit/allow. This here seems to me like a "I talked about lawn mowers to a friend, and just now I got a lawn mover ad, so my phone must be listening to me" situation.

010

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Honestly, I don't think Anthropic would do this. #altmanAI yeah, no doubt. The latter have been caught lying many, many times. There's an entire book about their boss being a pathological liar.

210

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

I agree LLMs are a weird, unelegant and quite stupid brute-force way to try to reach AGI/ASI. Nobody planned this. But here we are, and as a biologist I'm not that convinced that current models can't accelerate research so we end up with actual continual learning intelligence within the next years.

100

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 28/09/2026

Please give your definition of "singularity" and "aware". From what I can see, models don't need to be conscious to become powerful. Not sure that's what you mean by "aware". And recent models are not pure LLMs anymore.

200

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 27/09/2026

anthropic decided altmanAI should not be allowed to race to ASI, because that may kill us all. it might be argued the fight should look very different. but I haven't seen anybody else working harder and more honest to save our stupid asses.

000

cassandroid אַנטיפֿאַ @cassandroid.substack.com · 27/09/2026

It's more complex than that. altmanAI is amoral, hence the 25 million bribe (by brockman). musk is crazy and evil; he murdered millions of brown people. anthropic is truly fighting the good fight here, within the constrains of a fascist klepto-autocratic regime and late-stage capitalism.

100