iiinova — where your digital self keeps evolving

(btw, I tried to email, but won't be able to attend the Berkeley talk tomorrow -- great to meet you though, and hope to chat again!! :)

100

Alane Suhr @suhr.bsky.social · 2h

What a time, and what a meaning, to be around for the making ...

100

Alane Suhr @suhr.bsky.social · 2h

I really love this piece! "We don't necessarily feel that *we* know our beliefs are true: just that it is known, that someone else somewhere else has worked these things out in detail" I am always surprised how much more discourse there is about "AI" vs. about prediction markets ...

100

Alane Suhr @suhr.bsky.social · 3h

(if you know any scholars focusing on apotheosis I'd love to get names too!)

100

Alane Suhr @suhr.bsky.social · 3h

(I probably misinterpreted some hedging on your part in the QA that there *could* be some kind of "superhuman intelligence", that's what I was commenting on here in particular!) Anyway I am so glad you came to COLM, I really loved the talk, and it's so important for our community to hear!

110

Alane Suhr @suhr.bsky.social · 7h

Also completely disagree on the implication there is some thing "intelligence" which we can quantify and compare in some straightforward way!

110

Alane Suhr @suhr.bsky.social · 7h

I disagree that the Huggingface hack was not "part of OpenAI's plan" and I don't think it should matter whether it was!

110

Alane Suhr @suhr.bsky.social · 7h

+1 on Nissenbaum's point on "unauthorized (cultural) access" as a unifying aspect of the Navier-Stokes and Huggingface situation

120

Alane Suhr @suhr.bsky.social · 7h

And, as @adinawilliams.bsky.social mentioned to me, I would have loved to hear how these technologies are facilitating the destruction of the very communities they are "mechanizing"

120

Alane Suhr @suhr.bsky.social · 7h

Lovely talk from @himself.bsky.social at COLM! But I think the retort "the machine god is not coming" presupposes that there *could be* a situation in which there *is* a machine god, we just aren't there right now. I would love to hear more about the history of the cultural practice of apotheosis

270

Reposted by Alane Suhr

Hypervisible @hypervisible.blacksky.app · 04/10/2026

People ask whether these guys are true believers or just massively invested in the hype, and at the point where you are trying to lobby *the pope* to say you are creating the machine god, there is no meaningful distinction.

28815

Reposted by Alane Suhr

Tressie McMillan Cottom @tressiemcphd.bsky.social · 30/09/2026

Following the Pope’s encyclical about AI, you’re going to see a lot more religious intervention in shaping public opinion about tech companies. There are only two social institutions resilient enough to counter AI politically — states and religion. They bought the first. Standby on the second.

430975

Reposted by Alane Suhr

Mark Copelovitch @mcopelov.bsky.social · 03/10/2026

~Approximately no one in the general public knows because the leaders of the R1 research universities - whose budgets, programs, & personnel are entirely or significantly dependent on federal funding - have been totally absent from the public square in explaining & collectively protesting this.

27788308

Reposted by Alane Suhr

Hagen Blix @hagenblix.bsky.social · 03/10/2026

We need to rephrase the whole discussion in terms of hazardous products, instead of "minds". And then we need to make sure that there's liability. A model hacks someone? Put a C-suite guy in jail. A model produces sexual harassment material? Put a C-suite guy in jail. Watch how fast things get safe

12763258

Alane Suhr @suhr.bsky.social · 03/10/2026

why do we say "how do words get their meanings?"? why don't we say "how do meanings get their words?"

1352

Reposted by Alane Suhr

Anthony Moser @anthonymoser.com · 03/10/2026

a lot of people into ai want to talk about how it works, but they either mean the stack inside the software or the things it "can do now" that "can't be explained" but ai is a rhetorical technology. the way to understand how it works is to understand how rhetoric works

414933

Reposted by Alane Suhr

Anthony Moser @anthonymoser.com · 03/10/2026

but really, this post is just me reaffirming that I am not interested in whether they can do math because the primary function of LLMs is to help centralize, justify, and validate power, and they are doing that successfully already

312733

Reposted by Alane Suhr

Sumita Pahwa @sumitapahwa.bsky.social · 02/10/2026

"Claude has moral standing." Good, then Claude can go to hell.

2256

Alane Suhr @suhr.bsky.social · 01/10/2026

(also, this isn't new with the transformer architecture, infinite-precision RNNs are turing-complete; of course these are only theoretical, so there is some work on what happens with finite-precision RNNs, etc... arxiv.org/pdf/1805.04908)

arxiv.org

010

Alane Suhr @suhr.bsky.social · 01/10/2026

Watch out for the institutionalization of technologies!

010

Alane Suhr @suhr.bsky.social · 01/10/2026

my understanding is the transformer *architecture* is expressive enough as a formal object (like a formal language) to compute integer addition at arbitrary # of digits, but only if the weights are hand-set/programmed; training with SGD does not result in such weights

110

Alane Suhr @suhr.bsky.social · 01/10/2026

do you have the template for this? :) also, this is really not my area of expertise, but having vaguely lurked on this conversation, this paper may be of interest: arxiv.org/pdf/2410.18077

arxiv.org

210

Reposted by Alane Suhr

Justin Hendrix @justinhendrix.bsky.social · 29/09/2026

Plus, the "White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities" signed by Trump, Google's Pichai, Anthropic's Amodei, Meta's Zuckerberg, OpenAI's Brockman, XAI's Musk, Nvida's Huang.

In order to build a positive future for the American people and the world, we believe every company is responsible for developing its own technology safely and in a way that builds trust with customers and the public.

This starts with every company that is training and deploying frontier models having robust internal processes and controls to ensure that their technology behaves as intended and that any issues are promptly identified and resolved. Therefore, in addition to any other precautions, we believe each company should implement the following four layers of controls and audits:

(1) Implement robust internal controls to monitor the capabilities and alignment of its models during training and deployment around areas like cybersecurity, biosecurity, and chemical threats, and to ensure that its models do not hack or access technical systems in unintended ways.

(2) Empower an internal team to ensure all of the controls, monitoring, and detection are operating as intended, and that any issues are remediated.

(3) Partner with an independent external auditor or evaluator to carry out independent assessments of whether the controls, monitoring, and detection are operating as intended.

(4) Designate an independent committee of the board of directors to oversee and receive reports from the teams operating the controls and the internal and external auditors and evaluators, as well as to ensure any issues identified are remediated.

Together, these steps will give each company, its customers, and the public confidence that the technology is operating as intended.

The participating companies will meet regularly to establish standards and best practices to improve the safety of their systems.

Over time, it may make sense to codify these steps into laws or regulations. Regardless of whether this is required of companies, we believe that implementing these controls and audits is critical to ensuring a safe future for everyone, and each of our companies are ...

Page 2 (signature page)

[signature]
Donald J. Trump
President of the Unites States

[signature]
Sundar Pichai
Google

[signature]
Dario Amodei
Anthropic

[signature]
Mark Zuckerberg
Meta

[signature]
Greg Brockman
OpenAI

[signature]
Elon Musk
XAI

[signature]
Jensen Huang
Nvidia

3520749

Reposted by Alane Suhr

Rick CarBOOso’s Private Fire Crew @amandasmith.bsky.social · 01/02/2026

The only power we have in life is our presence. Money can buy a lot, but the real thing that legitimizes evil is if there are people that accommodate it. That’s the takeaway, for me, from all the people who Had No Choice but to beg for access to Epstein.

09512

Reposted by Alane Suhr

A. Feder Cooper @afedercooper.bsky.social · 27/09/2026

In general, I think disclosing COIs in science is a necessary feature of all papers, and that failing to do so calls into question the legitimacy of the work. I think it's especially important in this subfield because of the broader harm non-disclosure can potentially bring about in the real world.

101

Reposted by Alane Suhr

Tressie McMillan Cottom @tressiemcphd.bsky.social · 28/09/2026

Technology enthusiasts severely overestimate the average person’s desire or willingness or capacity to build and use AI agents. I understand some work roles that could mandate or encourage it but that is no way a tech revolution. Agents are best for structured, clearly delineated tasks.

331890304

Reposted by Alane Suhr

Annalee @flowerhorne.com · 26/09/2026

But seriously: Gifted Kid Syndrome will have you thinking that being The Smartest Person In The Room is your whole worth as a human. If AGI were real (it's not!), rationalists would have to accept that they are not The Smartest. This is why they consider it an existential threat.

021233

Reposted by Alane Suhr

ct @cteeze-nuts.bsky.social · 24/09/2026

I can't recommend highly enough this thread in which meta's muse model is shown to be so poorly hardened that it will tar its own VM filesystem and let you download it neuromatch.social/@jonny/11732...

neuromatch.social

jonny (nonvenomous) (@jonny@neuromatch.social)

Attached: 1 image Not only confirming this, its extremely easy. Almost no prompt injection resistance lmao

25626

Alane Suhr @suhr.bsky.social · 26/09/2026

You owe it to people who agree with you to make good arguments!!

010

Reposted by Alane Suhr

evacide @evacide.bsky.social · 26/09/2026

Listen, I just think it's weird that the same people who are saying they're worried that AI is going to wipe out humanity are also eager for military contracts that will give AI control of weapons.

431433341

Reposted by Alane Suhr

John Herrman @jwherrman.bsky.social · 23/09/2026

Excellent @himself.bsky.social post here, scratching at some of the overbearing tendencies — sometimes subtle and unintentional, sometimes open — in AI discourse www.programmablemutter.com/p/machine-go...

programmablemutter.com

Machine god metaphors eat your brain

There are other ways of thinking about AI

1308

Reposted by Alane Suhr

Hypervisible @hypervisible.blacksky.app · 21/09/2026

“My big fear with this technology is that people mistake it for something morally conscious, even though it flattens people into data representations and overlooks the depth of their lived complexity and dignity.”

news.yale.edu

How AI is changing the field of nursing

In a Q&A, nursing educator Christine Rodriguez explains where artificial intelligence belongs in health care — and where the role of actual human expertise is non-negotiable.

49922

Alane Suhr @suhr.bsky.social · 21/09/2026

yes we are planning to do this!!!

010

Alane Suhr @suhr.bsky.social · 21/09/2026

Yea I think there will be a lot of cool formats!! I will be experimenting with this at least a little bit over the year through this!! ai-and-society.github.io

ai-and-society.github.io

UC Berkeley AI & Society - AI & Society

AI & Society community at UC Berkeley

130

Alane Suhr @suhr.bsky.social · 21/09/2026

Ok ok ok so at the very least let's try to catch up sometime soon??? Also I gotta read my Garfinkel and Goffman before committing so I know at least vaguely how to get people to explain what they believe

020

Alane Suhr @suhr.bsky.social · 21/09/2026

Oh shit am I in my Maybe I Should Start a Podcast era

260

Alane Suhr @suhr.bsky.social · 21/09/2026

ummmmm yes I would love to do this??

230

Reposted by Alane Suhr

Anthony Moser @anthonymoser.com · 20/09/2026

This is why I say fascism is make believe at gunpoint

524266

Alane Suhr @suhr.bsky.social · 20/09/2026

as metaphors often do, it turns out that "AI" fits just as well into the category of "mimetically-fit cognitive virus", if you're looking through that lens at all! :)

020

Reposted by Alane Suhr

ve, the ghost of info past @vortexegg.com · 17/09/2026

This is what Trump's posting about Effective Altruism is about. Competition for whose fascist millenarian utopia should be the unifying vision for America's death drive into the abyss, in which they expect only the faithful will make it through whatever Singularity or Second Coming they imagine

0132

Reposted by Alane Suhr

ve, the ghost of info past @vortexegg.com · 17/09/2026

The thing people need to take seriously is that science fiction threats are also real threats when the people who hold power over you believe in them (though the nature of the actual threat is different)

47821

Alane Suhr @suhr.bsky.social · 17/09/2026

to clarify, I mean *from some people to other people (ok now I really need to go to a meeting)

000

Alane Suhr @suhr.bsky.social · 17/09/2026

I'm not sure if I've seen this elsewhere, but one perspective on all of this is that these arguments are about who should get which pieces of the pie, so to say, if pie pieces are like existing industries and communities these new companies can colonize

000

Alane Suhr @suhr.bsky.social · 17/09/2026

(through the development of better open-weight models, for example); "e/acc" people who want to maximize capital gains want cheaper models because it's a resource they are building some other product on. that is the argument)

100

Alane Suhr @suhr.bsky.social · 17/09/2026

(this is not to say that I am against regulation, rather the opposite. technologies designed to shift power away from people ought to be regulated. but pay attention to who has what stakes: "e/a" people want to "blanket ban" to maintain their ownership over a resource that is being devalued

200

Alane Suhr @suhr.bsky.social · 17/09/2026

to seriously take part in such an argument exhibits hubris as if one has some justified authority. notice how nowhere in these conversations are the *people* affected by developing technologies: medical researchers themselves, cybersecurity/infrastructure experts, regular people being surveiled

110

Alane Suhr @suhr.bsky.social · 17/09/2026

personally, I think that partaking in arguments about whether or not to blanket ban technological development is a waste of time and a willing engagement in a proxy argument that serves to distract from developing actual effective regulations (and other responses) that target real harms and risks

110