Подтвердите e-mail

Для публикаций, комментариев, реакций и сообщений подтвердите адрес.

Профиль

Jed Brown

Профиль Vively

Prof developing fast algorithms, reliable software, and healthy communities for computational science. Opinions my own. https://hachyderm.io/@jedbrown https://PhyPID.org | aspiring killjoy | refuse epistemicide | he/they

This is a good contextualization of the attempted smear. Unfortunately, with autotypers forging edit history, edit history will no longer be viewed as an affirmative defense. And political adversaries will weaponize doubt, making it impossible for those who are not longitudinal observers to discern.

Julia Serano

I hate this timeline & that I had to write this essay. I hate AI companies & LLMs & the randos who are now using "AI detectors" (also AI!) to smear authors they dislike: on Medium (friend link): juliaserano.medium.com/on-em-dashes... & Substack (no paywall): juliaserano.substack.com/p/on-em-dash...

011

10 GW of Rubin hardware (presumably what they would use) costs half a trillion dollars. Where are investors to believe that money comes from? NVIDIA circular financing to prop up their own stock?

020

From the paper, this "positive" just sounds like people soaking up the sycophancy (potentially to the chagrin of real people in their lives). I'm skeptical of the implied causality in the second pane (I suspect it's a coupled and reinforcing cycle). doi.org/10.1038/s415...

010

Yes, and the effect is huge. It's harder to measure, but I expect the influence bleeds over from personal use into maladaptive behavior toward other people. From simply being lower-quality friends to stalking and intimate partner violence.

Maggie Harrison Dupré

And here's our Feb story, which draws on 10+ individual cases we identified in which ChatGPT or another bot reinforced a user's delusional fixations on another person, leading to domestic abuse, harassment, and stalking: futurism.com/artificial-i...

150

Also, I think you'll agree that $750k-1M for a half-time sub-dean IS obscene. It's worse because the job duties infringe shared governance. (This salary is not the norm, but a bunch of people with not-this-high salaries approving it is a sign of something rotten.) diversityjobs.com/career/17211...

Senior Associate Dean for AI and Health Data Science job in Aurora, Colorado at University of Colorado Anschutz | Search Jobs Nearby or Remote JobsUniversity of Colorado Anschutz is hiring a Senior Associate Dean for AI and Health Data Science in Colorado. Learn more at DiversityJobs.com and apply today!diversityjobs.com
110

I think the problem with big salary discrepancy is that it feeds class solidarity with donors and against workers. That makes administrators worse stewards of the university mission. I don't think a university president, much less a dean, needs to make more than a governor or US cabinet secretary.

Jed Brown

Yes, and we need faculty unions in solidarity with students to fight that. The most celebrated university presidents from history got it right. (1939 article) Current presidents are too committed to class solidarity to voluntarily consider it. We need a credible threat. archive.is/202503240509...

110

Most would do a better job if they HATED schmoozing with donors and politicians, and took offense at attempts to coerce universities and attack academic freedom. The betrayal by AAU is a prime example, but it happens in smaller ways constantly. Betrayal of values, always time to work against faculty

030

These people deserve censure by AAU members and by faculty senates at their institutions. They are unfit to lead, and yet AAU carries disturbing influence with administrators of member universities.

AAUP

The Trump admin is back with version 2.0 of its higher ed compact. The reason is simple: The first one failed. Compact 2.0 lacks egregious elements of its predecessor. That’s a victory & affirms that the answer to authoritarianism is organized resistance—not compliance. www.aaup.org/news/new-com...

010

Not surveillance-based stalking, but still amplifies stalking and abuse.

Maggie Harrison Dupré

NEW: ChatGPT and other chatbots are feeding users' fixations on real people, reinforcing users' delusional ideas and obsessions about others as they descend into unwanted harassment, stalking, and/or domestic abuse — traumatizing victims and profoundly altering lives. futurism.com/artificial-i...

022

Yes. In retrospect, it was naive to hope that the AAU would not get back to their roots of carrying water for McCarthy-era censorious purges and Jim Crow.

Dominique Baker

Me on the AAU "The AAU claims to be comprised of America’s leading research unis...But if the member presidents choose the guise of institutional neutrality as a way to gain political cover, they may now be leading higher ed toward greater authoritarianism." www.insidehighered.com/opinion/colu...

022

OpenAI does not even pretend that $300B would be paid for by revenue. They'll need investors to accept over $200B in losses through 2029 without seeing any earnings until 2030 (if you believe the curve on a napkin that Altman is selling). the-decoder.com/openai-adds-...

010

200 million Americans, eh? Pretty outrageous. What does the cited source say? 67 million globally across a broader sector. How could that have happened?

110

Most cases will not be so clear-cut. Products already exist to forge edit history. Be it education, journalism, research, or creative works, the arms race to infer process from a context-free artifact is a losing game. All acceptable outcomes rely on trust. Not platitudes.

Jed Brown

That surveillance is an arms race with autotypers. The perceived necessity is tied to the centralization of traditional grades and credentialing. If you reject all of that, you have no choice but to build the mutual trust for students to voluntarily refrain despite no credible threat of enforcement.

120

German courts continue to lead the way in cutting through the hype to reasonably apply copyright law. Vendor indemnification clauses offer no protection.

Dan Martin

There is no AI without stealing intellectual property. That's the business model. Take everyone's work, build AI tools, dismantle access to knowledge, force people to pay to use AI. www.reuters.com/world/german...

041

There needn't be a contract or Anthropic association. His consulting work is paid by organizations that would look foolish without a solutionist hook. Refusal cannot be an option for a company selling responsible-washed agentic ✨ and hiring for "familiarity with AI/ML tools and prompt engineering."

Jed Brown

No AI-positive institution is deploying responsibly. If they say they are, it means they aren't checking or are silencing the people who are. Read Nik Suresh (hermit-tech.com/blog/ai-mani...) or listen to this interview, then read the lawsuit from the Mayo Clinic's director of research operations 👇

2123

As you note, communities can and do choose better, including addressing the many systemic harms beyond vehicular violence. But for that, communities need quality information. Jensen and cronies are once again opposed, and seeking government assistance to suppress.

Jed Brown

Will no one think of the poor billionaires? Over-leveraged on a bad product and fascism, celebrities don't even want their money, and respectable people with PhDs are teaching about harms and regulation. At least university administrators are upholding their duty by co-signing this gambling debt.

1107

Jensen "said cars were once portrayed as killing children, but the world changed its norms [..]". They really were killing extraordinary numbers of children, and remain a leading cause of death. Regulation (including liability and insurance) made a difference. Jensen and cronies fight that too.

Jacqueline Antonovich

Yesterday I was reading an academic book and learned that more than 500 children died in automobile accidents in New York City in 1914 alone, most after being struck by cars. I haven't been able to stop thinking about it. I tried to verify the number and found this poster from 1917.

2198

The same pattern is playing out at universities around the world. IT units parrot industry talking points oblivious (or defiant) of scholarship. Cushy platforms, speaking gigs, and political capital for those who push uncritical adoption, and abundant marginalizing tactics against critical scholars.

Olivia Guest · Ολίβια Γκεστ

Please see this as a general reminder that you're not alone. Many of us reject, refuse, repudiate artificial intelligence and the technology industry: Guest et al. (2026). Against the Uncritical Adoption of 'AI' Technologies in Academia. Digital Culture & Education doi.org/10.5281/zeno...

0227

Always tacitly (or explicitly) positioned such that increased use is the goal. More deployment, more industry partnerships. A whole institution with no resistance to risks (harms) or hype, no room for refusal. www.mayo.edu/research/dep...

1113
Показать ещё