BetterTec_h's profile picture. Concerned about the dark side of technology? Time spent here will be time well spent! By @jamesmart_in

BetterTech

@BetterTec_h

Concerned about the dark side of technology? Time spent here will be time well spent! By @jamesmart_in

Fascinating. 139 (so far) dignitaries - including former board member @hlntnr, who was instrumental in @sama’s failed ouster - sign a petition asking OpenAI to tell the truth. Good luck to them…

🚨 Breaking: A group of 100+ Nobel laureates, professors, whistleblowers, public figures, artists, and nonprofit organizations just released a letter asking OpenAI to tell the truth about its restructuring. Here’s what they had to say: 🧵

TheMidasProj's tweet image. 🚨 Breaking: A group of 100+ Nobel laureates, professors, whistleblowers, public figures, artists, and nonprofit organizations just released a letter asking OpenAI to tell the truth about its restructuring. 

Here’s what they had to say: 🧵
TheMidasProj's tweet image. 🚨 Breaking: A group of 100+ Nobel laureates, professors, whistleblowers, public figures, artists, and nonprofit organizations just released a letter asking OpenAI to tell the truth about its restructuring. 

Here’s what they had to say: 🧵


2x as many kids in the UK use ChatGPT today than just two years ago. More vulnerable kids are 4x more likely to become friends with AI companions. The tragic suicide of Sewell Setzer, caused by Character AI, was not a one-off. More in our latest blogpost: bettertech.blog/2025/07/19/cou…


Some not only have AI romantic partners: one in this incredible article has several. & USES CHATGPT FOR ADVICE on how to handle each one. AI’s environmental impact is scary. Its social impacts vs the loneliness epidemic could be even more devastating 😨😱 wired.com/story/couples-…


Have we reached a tipping point when it comes to the impact of AI? The mood at traditionally techno-solutionist event @VivaTech was positive re. our prospects of switching to more #frugalAI. Our report, featuring our panel with @SashaMTL & more: bettertech.blog/2025/06/13/viv…

BetterTec_h's tweet image. Have we reached a tipping point when it comes to the impact of AI? The mood at traditionally techno-solutionist event @VivaTech was positive re. our prospects of switching to more #frugalAI. Our report, featuring our panel with @SashaMTL & more: bettertech.blog/2025/06/13/viv…

In which we discover that, by 2030, the biggest AI supercomputers will demand THIRTY times more energy than they do today. And this is coming from @EpochAIResearch, who were last seen minimising the impact of 1 ChatGPT prompt #thisisfine arxiv.org/pdf/2504.16026 via @SashaMTL


On Character AI, you can now create your AI companion which looks however you want, including by uploading photos. As a vague privacy protection move, real people are made into sort-of lookalikes. But still, be afraid. Be very afraid youtu.be/LEPaWVbRJ1E?si… cc @HumaneTech_

BetterTec_h's tweet card. Introducing AvatarFX

youtube.com

YouTube

Introducing AvatarFX


"ChatGPT gets things wrong. It makes up references. And it’s that system that’s going to be deciding whether you get Medicaid or not?" @carolecadwalla nails exactly why tech bros+Trump is a match made in hell. Thanks Carole for encouraging us to fight ✊🏻 youtu.be/vG7CvbccdVM?si…

BetterTec_h's tweet card. Carole Cadwalladr - Broligarchs, AI, and a Techno-Authoritarian...

youtube.com

YouTube

Carole Cadwalladr - Broligarchs, AI, and a Techno-Authoritarian...


To cut costs and further bend the knee to POTUS, WaPo is introducing an AI editor with a "story strength tracker". 🤢 As @carolecadwalla just pointed out, this sort of AI investment may well come from Saudi ruler MBS, who famously got a WP journo killed mediaite.com/media/print/wa…


Delighted to feature in my first ever podcast, especially in one that’s taught me so much about #greenIT. In a nutshell: if it’s not measurably better for the planet than its negative impacts, it’s not "AI for good"; it’s greenwashing podcasts.apple.com/fr/podcast/env…


BetterTech أعاد

Yes, "green prompting" is a thing! Re. this white paper, and confirmed by other experts, 2x shorter LLM results - e.g. by adding "be concise" to your prompts - uses 2x less energy (because it can half the no of tokens generated). Nice! arxiv.org/abs/2503.10666 cc @DataForGood_FR


The ever-incredible @doctorow finally nails Careless People - and therefore Facebook - in just over 100 tweets. A must-read testament to why it’s such an important book 👏🏻👌🏻👏🏻

I never would have read *Careless People*, Sarah Wynn-Williams's tell-all memoir about her years running global policy for Facebook, but then Meta's lawyer tried to get the book suppressed and secured an injunction to prevent her from promoting it: npr.org/2025/03/14/nx-… 1/

doctorow's tweet image. I never would have read *Careless People*, Sarah Wynn-Williams's tell-all memoir about her years running global policy for Facebook, but then Meta's lawyer tried to get the book suppressed and secured an injunction to prevent her from promoting it:

npr.org/2025/03/14/nx-…

1/


AI is going great, part 67: customer support bot makes up a policy that doesn’t exist. Whilst the company in question apologised this time (unlike Air Canada), GenAI’s tendency to invent rather than admit it doesn’t know stuff is increasingly problematic flip.it/c1Nr-D


Loading...

Something went wrong.


Something went wrong.