Rebecca Ackermann's Avatar

Rebecca Ackermann

@rebackermann

Writing about art, culture & tech for MIT Tech Review, Vox, Esquire & more | I have eclectic interests

418
Followers
320
Following
44
Posts
10.07.2023
Joined
Posts Following

Latest posts by Rebecca Ackermann @rebackermann

Preview
Anthropic Drops Flagship Safety Pledge In an abrupt shift, the companyΒ may release future AI models without ironclad safety guarantees

BuT CaN tHe RoBoT tHiNk?

time.com/7380854/excl...

25.02.2026 19:18 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

one of many reasons that the organized critique and refusal movement is overwhelmingly led by women (yes including women computer scientists, cognitive scientists, and engineers)

27.12.2025 12:36 πŸ‘ 525 πŸ” 132 πŸ’¬ 4 πŸ“Œ 6

Yes, except for that breathless EA piece awhile back…

24.02.2026 14:25 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
Imagine you have two machines. One you can open up and examine all of its workings, and if you give it every picture of a cat on the whole internet, it can reliably distinguish cats trom non-cats. The other is a black box and it can also reliably distinguish cats from non-cats if you give it half a dozen pictures of cats, some apple sauce, and a hug. These machines sort of do the same thing, but even without knowing how the second one works 1 am extremely confident in saying it doesn't work the same way as the first one.

Imagine you have two machines. One you can open up and examine all of its workings, and if you give it every picture of a cat on the whole internet, it can reliably distinguish cats trom non-cats. The other is a black box and it can also reliably distinguish cats from non-cats if you give it half a dozen pictures of cats, some apple sauce, and a hug. These machines sort of do the same thing, but even without knowing how the second one works 1 am extremely confident in saying it doesn't work the same way as the first one.

A.I Isn’t People www.todayintabs.com/p/a-i-isn-t-...

23.02.2026 22:42 πŸ‘ 375 πŸ” 106 πŸ’¬ 10 πŸ“Œ 31

Plus! The computer scientist from Brown also worked for Google DeepMind!! But that’s not disclosed, she’s just an academic asking questions!

13.02.2026 22:26 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

You know it’s going to be bad when you can hear the PR email pitch in the subhed.

13.02.2026 21:21 πŸ‘ 9 πŸ” 2 πŸ’¬ 1 πŸ“Œ 0

I am so tired of all the things that we said would happen, and were routinely ignored or dismissed by tech bros and neoliberals for saying would happen, continuing to happen, in exactly the way that we said they would happen.

30.01.2026 00:16 πŸ‘ 232 πŸ” 81 πŸ’¬ 1 πŸ“Œ 4

Hard no to this: French publisher Harlequin has told translators that their work will now be done by AI. They can β€˜revise’ AI translations at much lower rate. It’s a subsidiary of Harper Collins so treat this as a test for rollout elsewhere

18.12.2025 21:48 πŸ‘ 516 πŸ” 307 πŸ’¬ 22 πŸ“Œ 62

The least of anyone’s problems but wow the iPhone β€œLiquid Glass” update is just more confirmation that everything now is worse for no good reason.

19.12.2025 20:12 πŸ‘ 2 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0

Write about this pls (or I will!!)

15.12.2025 18:47 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
Preview
Why Waymo can't compete with the bus Public infrastructure deserves to win out over private convenience.

Hell yes @rebackermann.bsky.social:

"If we don’t want to live alongside other city residents, why do we live in this city at all?"

www.sfexaminer.com/forum/waymo-...

15.12.2025 17:40 πŸ‘ 8 πŸ” 4 πŸ’¬ 2 πŸ“Œ 0

❀️🚍❀️

15.12.2025 17:42 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Remember that Intuit (the owner of Turbotax) lobbied extensively to kill Direct File, which did tax filing for free and somehow managed to make a good and intuitive user experience without AI.

18.11.2025 17:27 πŸ‘ 518 πŸ” 228 πŸ’¬ 9 πŸ“Œ 4
Preview
The Publishing Industry’s Most Swoon-Worthy Genre Tight-knit but open-armed fans have made romance an especially hot commodity.

What’s driving Romancelandia’s success, and can other genres successfully emulate it? Spoiler alert: There’s a phenomenon called β€œan open-elite network.” @rebackermann.bsky.social for @theatlantic.com www.theatlantic.com/books/2025/1...

14.10.2025 19:41 πŸ‘ 1 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0
Preview
Romance Fiction’s Secret Weapon Tight-knit but open-armed fans have made it an especially hot genre.

The tight-knit but open-armed community that supports romance novels has become unignorable, enviableβ€”and exceedingly difficult to replicate, Rebecca Ackermann writes.

07.10.2025 17:15 πŸ‘ 13 πŸ” 3 πŸ’¬ 0 πŸ“Œ 0
Business Insider headline saying: "Sam Altman says your ChatGPT therapy session might not stay private in a lawsuit
By Robert Scammell"

Business Insider headline saying: "Sam Altman says your ChatGPT therapy session might not stay private in a lawsuit By Robert Scammell"

Wired headline that says "ICE Is Getting Unprecedented Access to Medicaid Data
A new agreement viewed by WIRED gives ICE direct access to a federal database containing sensitive medical data on tens of millions of Americans, with the goal of locating immigrants."

Wired headline that says "ICE Is Getting Unprecedented Access to Medicaid Data A new agreement viewed by WIRED gives ICE direct access to a federal database containing sensitive medical data on tens of millions of Americans, with the goal of locating immigrants."

25.07.2025 14:52 πŸ‘ 537 πŸ” 160 πŸ’¬ 7 πŸ“Œ 19

I’m still not over the fact that there was an extremely popular award-winning movie over a decade ago about the perils of falling in love with a chatbot and a company decided to make the chatbot in the image of the movie’s chatbot and now quite a few people have decided to date it

23.06.2025 04:47 πŸ‘ 5461 πŸ” 702 πŸ’¬ 118 πŸ“Œ 76
Preview
Analysis | AI firms say they can’t respect copyright. These researchers tried. A new effort using only openly licensed data may have implications on thorny policy disputes around copyright and AI

A lot of people say generative AI shouldn't infringe on copyright. These researchers actually tried to do it. The result: an 8 terabyte dataset of text that's openly licensed or in the public domain & 7 B parameter model that performs as well as Meta's Llama 7B www.washingtonpost.com/politics/202...

05.06.2025 17:44 πŸ‘ 766 πŸ” 239 πŸ’¬ 15 πŸ“Œ 36

Ughhhh I’m sorry (to you & to all of us for losing you)

20.05.2025 16:16 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

What no!!

20.05.2025 16:13 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
Preview
Bay Area NEA Grantees

I started a spreadsheet of Bay Area NEA Grantees for FY23–FY25. So far, there's ~85 arts orgs whose funding was likely withdrawn without any reimbursements going out. I'm expecting 3x that number when I get to the terminated grants.

docs.google.com/spreadsheets...

07.05.2025 01:37 πŸ‘ 17 πŸ” 7 πŸ’¬ 2 πŸ“Œ 2
Preview
How AI is interacting with our creative human processes Three books examine what we gain and lose when we let machines create.

Gift link! ter.li/RAckermann

14.04.2025 17:07 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
The rapid proliferation of AI in our lives introduces new challenges around authorship, authenticity, and ethics i. Work and art. How can we make sense of these machines, not just use them?

The rapid proliferation of AI in our lives introduces new challenges around authorship, authenticity, and ethics i. Work and art. How can we make sense of these machines, not just use them?

For @technologyreview.com I wrote about how the names we give generative machines really matter.

14.04.2025 17:06 πŸ‘ 3 πŸ” 1 πŸ’¬ 1 πŸ“Œ 0

Such a fascinating read and so fun to write about!

12.04.2025 21:12 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Preview
How AI is interacting with our creative human processes Three books examine what we gain and lose when we let machines create.

I wrote about what we gain & lose when we let machines make art for us (& about @vauhinivara.bsky.social’s eerie new book) www.technologyreview.com/2025/04/11/1...

12.04.2025 13:51 πŸ‘ 7 πŸ” 3 πŸ’¬ 1 πŸ“Œ 0

Calling it now: they’ll launch stablecoins as the new war bonds

09.04.2025 15:31 πŸ‘ 2 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0
Preview
OpenAI declares AI race β€œover” if training on copyrighted works isn’t fair use National security hinges on unfettered access to AI training data, OpenAI says.

So it’s over then, works for me

13.03.2025 20:20 πŸ‘ 8587 πŸ” 1623 πŸ’¬ 348 πŸ“Œ 729
Preview
Is AI the Bitter Endβ€”or the Lucrative Futureβ€”of Book Publishing? As the law fights to catch up to Big Tech, the future of books hangs in the balance. Are writers doomed by β€œthe biggest rip-off in creative history” or could AI offer new ways of making a living?

An AI model that can write you say? Wonder who it learned how to do that from (& if these fair use cases are ever going to get settled) www.esquire.com/entertainmen...

13.03.2025 23:24 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

It is now

27.02.2025 21:29 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Preview
Your boss is watching Monitoring technology is increasing the power imbalance between companies and workers. Protections lag far behind.

More on how electronic monitoring is widening the dangerous power gap between workers & employers across the country www.technologyreview.com/2025/02/24/1...

25.02.2025 21:05 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0