
StructuredStories
@StructStories
Followers
2K
Following
718
Media
272
Statuses
3K
News should be economically sustainable, differentiated, adaptable, personally relevant and truly accessible to everybody. Computers can help.
London, England
Joined October 2013
How will AI reshape the information ecosystem?. Earlier in 2024 I worked with Shuwei Fang & @OpenSociety to apply scenario planning to that question, collecting scenarios from ~900 people & testing them in a 60-person workshop. This report is the result.
opensocietyfoundations.org
The AI in Journalism Futures project seeks to broaden and deepen the conversation about AI’s potential impact on journalism and civic information around the world
2
1
13
RT @aryehazan: as opposed to human brains, which have direct access to Platonic reservoirs of novel truths.
0
24
0
So far @DKokotajlo's AI 2027 is right on track . .
New paper & surprising result. LLMs transmit traits to other models via hidden signals in data. Datasets consisting only of 3-digit numbers can transmit a love for owls, or evil tendencies. 🧵
0
0
1
RT @jonathanstray: I'm an applied epistemologist. I try to understand how the philosophy of knowing things translates into truth-seeking in….
0
1
0
20 yrs ago I was stunned by @JeffCHawkins book ‘On Intelligence’ & its hierarchical AI. To me it was a credible path to AGI – more plausible than AI ‘emerging’ from NNs. This new work from @Sapient_Int shows fascinating results from a similar approach.
1
1
6
Some deep culture shifts behind this.
At @Google, we are moving from a writing‑first culture to a building‑first one. Writing was a proxy for clear thinking, optimized for scarce eng resources and long dev cycles - you had to get it right before you built. Now, when time to vibe-code prototype ≈ time to write PRD,.
0
0
0
It certainly seems that we're on the way to Google Zero, or at least to Google 'so-little-traffic-that-it-doesn't-really-matter'.
pewresearch.org
In a March 2025 analysis, Google users who encountered an AI summary were less likely to click on links to other websites than users who did not see one.
0
0
1
RT @DKThomp: Yes. Writing is not a second thing that happens after thinking. The act of writing is an act of thinking. Writing *is* think….
0
7K
0
An OpenAI researcher building anticipation for GPT-5. This will be a big moment for AI, either way. If GPT5 is kinda like Grok4 that will suggest we've maybe plateaued a bit. If it's a big step forward then all bets are off and we'll all need to reset. Again. .
0
0
0
RT @yoheinakajima: oh, you’re into AI? then tell me about…. GPT, Claude, Gemini, LLaMA, Mistral, Mixtral, Grok, DeepSeek.Kimi, Qwen, Ernie,….
0
226
0
RT @cIass_man: there's a caste in american society that consumes as much blue-chip media (nyt, new yorker, npr) as possible and then quizze….
0
524
0
Covid constituted a “stress test” that “the central truth-seeking departments of liberal democracy: journalism, science & universities” frequently flunked. Trust is earned. Or not. (see also Tapper & Thompson's 'Original Sin').
washingtonpost.com
A new book on the pandemic response dissects how U.S. institutions badly failed the test.
0
0
0
Working with experts to achieve goals is hard.
Unless and until agents really do work at expert level, the benefits of AI use are going to be contingent on the skills of the AI user, the jagged abilities of the AI you use, the process into which you integrate use, the experience you have with the AI system & the task itself.
0
0
0
The insane rate of progress in AI often overshadows the even more insane rate of progress in deployed solar and wind. The lack of media coverage of this profound revolution in energy is truly weird, but @billmckibben's review here is authoritative.
newyorker.com
In the past two years, without much notice, solar power has begun to truly transform the world’s energy system.
0
1
1
Can y'all maybe just gimmie a minute to figure out Dia and Comet first?.🙄.
reuters.com
The browser is slated to launch in the coming weeks and aims to use artificial intelligence to fundamentally change how consumers browse the web.
0
0
1
AI 2027 is still the most plausible description of the potential emergence of AGI I've seen, but the AI Futures team says more likely AI 2029 or AI 2031. I also think they may under-estimate diffusion friction, even for AGI. @sama's "we get AGI & nothing changes" rings truer. .
@jonst0kes So it seems you think we are wrong about timelines. That's reasonable. I'm currently at 50% by end of 2028, others on the team are a few years later. What are your timelines? Also, I assume you think we are wrong about more than just timelines, otherwise you wouldn't answer at.
0
0
1
I'm seeing this a lot at news organizations. Also, 'human-in-the-loop' isn't as absolute as it was a year or two ago. .
Axios is changing its editorial policies to be more friendly to AI use. In a recent note to staff, Axios said it was revising its AI language. "The new language replaces our original framing which was centered on how everything we do 'will be written or produced by a real.
0
0
2