Sign in

Matthew Chalmers

@matthewchalmers.bsky.social
227 followers 59 following 305 posts

Computer scientist into Ubicomp, HCI, theory and (a long time ago) data visualisation. Also kind of keen on mountain things, fine food things, and fine food in the mountains.

PostsRepliesMedia
Reposted by Matthew Chalmers
Ed Zitron @edzitron.com · 5h
42986168
Reposted by Matthew Chalmers
Dr Abeba Birhane @abeba.blacksky.app · 5h
we have a policy for the use of genAI at the @aial.ie. we reflect on values key to scientific integrity, such as replicability, and err on the side of not using it blogpost: aial.ie/blog/gen-ai-... pdf: zenodo.org/records/2322...
A flowchart that guides reflective thinking around generative AI in the context of research. It highlights key, interconnected values underpinning scientific integrity, alongside practical and wider socio-political considerations when deciding whether to use generative AI for research.
214665
Matthew Chalmers @matthewchalmers.bsky.social · 10h
"By June 29th, Switzerland [...] reached 'Glacier Loss Day' – [...] every further litre of meltwater comes directly from the ice. [Abstract] until you're standing on a glacier at 3am, trying to work out whether the snow in front of you is a bridge or a trapdoor" www.ukhillwalking.com/articles/fea...
ukhillwalking.com
Alpine Climbing in a Warming World - What does the future hold?
Following the catastrophic flooding in Nepal and Tibet, the effects of climate change in the mountains are becoming ever more obvious. The warming of the Alps is having a huge impact on mountain...
000
Matthew Chalmers @matthewchalmers.bsky.social · 12h
We need basic rules about data centres: ⚡ 100% clean power 💧 Zero water waste 🗣️ Local community consent Sign Foxglove's petition to stop data centres harming our communities and environment now. you.38degrees.org.uk/petitions/ba...
you.38degrees.org.uk
Ban data centres that harm our communities and climate
Big Tech’s giant data centres swallow as much power as millions of homes, guzzle our water, and risk driving up our energy bills - all with zero limits. We need basic rules: ⚡ 100% clean power 💧 Ze...
011
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 13h
That aside: the idea open models are "better" in terms of sustainability is wildly overplayed. There's *some* difference but it's small Ecosia are scamming a huge, huge number of eco-minded people and pushing them into wasteful, private-jet-esque search habits cleerdash.sustainableaigroup.com
a scatter plot showing open vs closed makes no difference to climate impacts
46623
Reposted by Matthew Chalmers
Butlerian Jihadist of House Dornick @shimminykricket.blacksky.app · 21h
Harvard just dropped a study AI productivity in software engineering and it found the average increase in productivity from using Claude code is.....zero and for exactly the reasons you'd think fion.ac/jellyfish.pdf
9056512226
Reposted by Matthew Chalmers
Johnny Ryan @johnnyryan.bsky.social · 15h
This looks like an extraordinary capitulation by the EU. Bravo to France for holding fast. www.theguardian.com/us-news/2026... @disinfo.eu @kajakallas.bsky.social
theguardian.com
EU drops sponsorship of disinformation conference at Trump officials’ request
Lithuania and Canada were also pressured to withdraw, and days before #Disinfo2026, logos were wiped off website
55113
Reposted by Matthew Chalmers
absolute horses @jcoglan.com · 07/10/2026
I once had someone tell me, and I do mean this literally, that LLMs transcend the limits of turing machines and computability because they are not constrained by having to give correct answers
812737
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 06/10/2026
This is a good illustration of how so much of AI theft isn't indirect (ie, data sucked into a model to train the chatbot) - it's actually **direct** copying, but *enabled* by GenAI. Of course it is: chatbots absolutely suck at creating good, new stuff.
25950422
Matthew Chalmers @matthewchalmers.bsky.social · 07/10/2026
"[S]ometimes what seems to be a decision about convenience can be read as a choice about values. When businesses could have commissioned a designer, using AI to produce a poster can be seen as opting out of the local creative community." observer.co.uk/news/technol...
observer.co.uk
AI images are easy to make, but one Bristol pub knows the...
Small businesses are using the tech to create inexpensive wall art, flyers and social media posts, but there’s a growing public backlash over ‘values’
000
Matthew Chalmers @matthewchalmers.bsky.social · 07/10/2026
www.theguardian.com/technology/2...
theguardian.com
Google told to halt work on datacentres in Finland over environmental concerns
Order on two construction sites comes after Google failed to carry out environmental impact assessments
000
Matthew Chalmers @matthewchalmers.bsky.social · 07/10/2026
www.theguardian.com/technology/2...
theguardian.com
Is regulation enough to stop AI systems killing us all? | Letters
Letter: Martyn Thomas wonders how any rational person can argue that independent oversight and regulation will be sufficient
000
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 06/10/2026
Simon Stålenhag is the best www.instagram.com/p/C3j9F6kq6p5/
simon_stalenhag's profile picture
simon_stalenhag

6,191 likes
simon_stalenhag's profile picture
simon_stalenhag
The way ai is marketed is just a new version of the old promise of perpetual motion. It cannot produce any new information about the world that hasn’t already been discovered by humans. A generated video of an ant tells us nothing new about ants that wasn’t already in the original data. I always considered the creative process as a form of discovery. The promise of a perpetual motion machine of creativity rests on very muddy assumptions of how new ideas are generated. Knowledge and creativity is not a library of data, but a continuous, dynamic process that requires interaction with the real world.

Just as with perpetual motion machines, artificial content might look like free energy for a while, but eventually it will grind to a halt unless real humans makes the effort to go out there and wind it up again.
29532
Matthew Chalmers @matthewchalmers.bsky.social · 06/10/2026
www.theguardian.com/uk-news/2026...
theguardian.com
Drax datacentre could create nearly twice the yearly emissions of all Gatwick flights
Exclusive: Campaign group’s analysis of North Yorkshire plans finds datacentre would have to burn 4.9m tonnes of wood a year
000
Matthew Chalmers @matthewchalmers.bsky.social · 06/10/2026
Summary/overview here: cepr.org/voxeu/column...
cepr.org
000
Matthew Chalmers @matthewchalmers.bsky.social · 06/10/2026
Using the AI tool raised the quality of the drafts and compressed performance differentials, especially among junior lawyers. [On] a subsequent task without AI tools, expertise gains from the previous AI use appeared only among senior lawyers with seasoned judgement papers.ssrn.com/sol3/papers....
papers.ssrn.com
Does AI Assistance Enhance or Erode Expertise? Evidence from a Three-Month Field Experiment in Patent Drafting
Whether AI assistance builds or erodes professional expertise is unsettled. In a pre-registered three-month randomized controlled trial, we gave 133 practicing
200
Reposted by Matthew Chalmers
Benjamin Riley @benjaminjriley.bsky.social · 05/10/2026
I'm again grateful to @theverge.com for publishing my newest long-form essay on the limits of the computational model of the mind, the benefits of understanding both biological and cultural evolution to explain human thinking, and the danger of cognitive hot dogs in education.
theverge.com
Our minds aren’t equipped to handle AI
AI is junk food for the mind: easy, tempting and ultimately very bad for you.
719548
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 06/10/2026
The LOVELY people at Klimakultur had me in to write about the unfathomable amount of death and destruction that can still be avoided if Norway + Equinor decide to allow a wind down of fossil fuel production A lot to lose here....but we are already saving lives!!! www.klimakultur.no/nyheter/esti...
18424
Matthew Chalmers @matthewchalmers.bsky.social · 06/10/2026
www.theguardian.com/technology/2...
theguardian.com
‘Pull the plug’: protesters resort to direct action against AI firms
Campaign groups report surge in membership after a spate of AI safety alerts and apocalyptic warnings
000
Reposted by Matthew Chalmers
404 Media @404media.co · 05/10/2026
“Indiscriminate mass surveillance," a Meta Muse scoop, and more in today's daily newsletter.
404media.co
Muse escapes containment
“Indiscriminate mass surveillance," a Meta Muse scoop, and more in today's daily newsletter.
1014563
Reposted by Matthew Chalmers
Nicholas Slayton @nslayton.bsky.social · 04/10/2026
What. What?! WHAT?!?! futurism.com/artificial-i...
futurism.com
Trump Was Reportedly Pushed Into Invading Venezuela by Conversation With Grok
Trump reportedly consulted with Grok for hours on his next course of action in Venezuela before launching a military invasion.
40510156
Reposted by Matthew Chalmers
Carl T. Bergstrom @carlbergstrom.com · 03/10/2026
Columnist Stephen Rattner probably got screwed over by the headline writer, but I'm seriously floored at how poor a job these six charts do at showing that we need AI.
nytimes.com
Opinion | Six Charts That Show Just How Much We Need A.I. (Gift Article)
All the worry about artificial intelligence should not obscure this key point: If we don’t develop it, we’ll be a lot worse off.
2137471
Reposted by Matthew Chalmers
heute jedoch nicht @heutejedochnicht.bsky.social · 12/09/2026
liebe
Meta-Adbusting in der Londoner U-Bahn
Jeffrey Epstein mit Ray Ban Meta-Smartglasses und Denkblase „Finally, a pair that suits me...“
Caption: „GLASSES FOR PEOPLE WHO DON'T DO CONSENT“
81566423
Reposted by Matthew Chalmers
Wolfie Christl @wchr.bsky.social · 02/10/2026
The web and mobile apps of major conversational AI engines transmit prompts, titles and conversation permalinks to advertising data firms, alongside device IDs, hashed email and other persistent identifiers. Comprehensive analysis, like always from this team: jorgegarciaherrero.com/wp-content/i...
26667
Reposted by Matthew Chalmers
Hisham Zerriffi @hishamzerriffi.bsky.social · 02/10/2026
Pew tested AI synthetic surveys against actual survey data. Conclusion: “This experiment taught us that, at this time, AI models are not an adequate replacement for traditional polling on topics of broad public importance.” This is an all time bad idea. No need to hedge with “at this time.”
pewresearch.org
Can AI Stand In for Human Survey-Takers? Not Really
Synthetic surveys using digital twins do a poor job estimating the public mood and are especially error-prone for some subgroups, including Republicans and Black adults.
1123680
Reposted by Matthew Chalmers
Dr Abeba Birhane @abeba.blacksky.app · 02/10/2026
this is why activism (for the preservation of fundamental rights) has become costly
0509
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 02/10/2026
How on Earth did I not read this yet? really interesting and honestly kind of brazen modelling of the entire lifecycle - downstream, data centres, upstream use - of AI. Puts AI industry at 1.1% of total global energy, about the same for global emissions. More than 10x previously stated estimates
Final Results and Concluding Thoughts
Adding it all up, we estimate that the AI sector is responsible for almost 1,900 TWh of global energy consumption in 2026. And after accounting for additional factors that we did not explicitly analyze, like the construction of new electrical substations, robotic-centric warehouses, water reclamation facilities, and other infrastructure specifically designed to support the data center buildout, it’s likely that total AI-driven global energy demand is already closer to 2,000 TWh. Even at 1,900 TWh, the AI share of total global energy use in 2026 would be roughly 1.1%. We also estimate, after adding up all the sectoral numbers and landing at roughly 650 million metric tons of AI-driven CO2 emissions, that the AI share of global GHG emissions in 2026 stands at about 1.2%, with total annual global emissions sitting at roughly 55 gigatons of CO2-eq, albeit with a wide range of figures from EDGAR to UNEP (anywhere from 53 gigatons to 58 gigatons depending on a broad array of different assumptions). As we’ve discussed already, these numbers are much higher than many others in the media and academia have guessed on the basis of simplistic calculations and naïve assumptions.

The projections up to 2030 are obviously far worse and imply a catastrophic impact to global ecological stability. ChatGPT alone “processes approximately 2.5 billion prompts per day worldwide.”122 The scale of AI growth and diffusion globally is truly astounding, and the industry’s corresponding ecological footprint will soon rival that of major countries like India. It's not farfetched at all to think that the AI share of total global energy demand could rise to something like 5% by 2030, rivaling and in some cases surpassing the footprint of established dominant industries. The potential consequences of this extra resource pressure on biodiversity, global warming, freshwater scarcity, and biogeochemical cycles could be absolutely staggering and devastating for the…
311557
Reposted by Matthew Chalmers
Carl T. Bergstrom @carlbergstrom.com · 19/09/2026
When we talk about the costs that LLMs impose on society we should not forget the fact that nothing works anymore because everyone is so scared of getting their shit scraped. I can't even use google scholar from my university or starlink because they send too much traffic.
Google	
Sorry...
We're sorry...

... but your computer or network may be sending automated queries. To protect our users, we can't process your request right now.
See Google Help for more information.
984459968
Reposted by Matthew Chalmers
Katie Mack @astrokatie.com · 13/02/2025
I am no longer accepting the argument "there are bigger problems than this!" unless you are personally working to stop the ultimate Heat Death of the Universe.
337187182159
Reposted by Matthew Chalmers
Ed Zitron @edzitron.com · 02/10/2026
This is the dodgiest and most-desperate thing I’ve seen in the bubble so far. Amazon selling their chips to a group of investors and leasing them back - an off balance sheet arrangement done in plain sight. Really scraping the bottom of the barrel here. And they’ll probably buy more!
Amazon is seeking to offload about $8bn of advanced Nvidia chips to external investors through a new vehicle aimed at strengthening its balance sheet, according to people familiar with the matter.
The Seattle-based cloud giant has held talks with investors in recent weeks to gauge interest in the deal, which would allow the company to spin off thousands of Grace Blackwell chips it is deploying in data centres across the US into a special-purpose vehicle.
Amazon will then lease the advanced AI chips back from the SPV, which would tap outside investors through debt issuance. By unloading the expensive semiconductors to investors, the company could adopt a more asset-light approach to its balance sheet.
611112264
Reposted by Matthew Chalmers
Emily M. Bender @emilymbender.bsky.social · 01/10/2026
New from @timnitgebru.blacksky.app and me: "It is high time for us to regulate usage of systems sold as “AI”, not because they are magical machines, but because they are error-prone products that shouldn’t be used in high stakes scenarios." www.theguardian.com/commentisfre...
theguardian.com
Forget ‘superintelligence’: error-prone AI nearly sparked world war three this month | Timnit Gebru and Emily M Bender
The risks of AI aren’t what we think they are, as a recent security incident between China and the United States reveals
13974417
Matthew Chalmers @matthewchalmers.bsky.social · 01/10/2026
“[Thylmann] was once the man behind the pornography sites Pornhub, YouPorn, RedTube and several other... He managed to buy all of those thanks to software he developed in the 1990s, which helped website operators track the ads and links users clicked on.” observer.co.uk/culture/book...
observer.co.uk
How porn invented the attention economy | The Observer
In Guilty Pleasure, Patricia Nilsson reveals how the adult industry commodified intimacy and emotion – and shaped the internet in its image
000
Reposted by Matthew Chalmers
404 Media @404media.co · 30/09/2026
"Do you watch the news? Do you listen to the radio? Do you read anything about what's going on in the world?” one of the judges asked the lawyer, who submitted false testimony and witnesses after using ChatGPT.
404media.co
Lawyer Cites ChatGPT-Invented Fake Witnesses in Murder Appeal
"Do you watch the news? Do you listen to the radio? Do you read anything about what's going on in the world?” one of the judges asked the lawyer, who submitted false testimony and witnesses after using...
920370
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 28/09/2026
Another day, another data centres ❤️ fossil fuels story. In this absolutely bloody amazing article, a company details how they can install data centres directly on top of gas pipelines. Still waiting for AI to solve climate change www.datacenterdynamics.com/en/analysis/...
Sapphire Technologies thinks so and is proposing to work with another company, Anax Power, to install modular data centers on the gas network, using pipeline pressure to provide cooling, as well as power, to the facilities.
24411
Reposted by Matthew Chalmers
Johnny Ryan @johnnyryan.bsky.social · 28/09/2026
Important breakthrough in Europe's icebreaker Big Tech class action opens the path for class actions to hold Big Tech accountable using the GDPR. ICCL v Microsoft class action update... Full decision in link --> enforce.ie/digital-data...
enforce.ie
ICCL secures landmark precedent against Big Tech - Enforce
Class action will target “Real-Time Bidding” (RTB) data breach in Microsoft’s advertising system, and is anticipated to affect Microsoft’s operations across the EU
22728
Reposted by Matthew Chalmers
Carl T. Bergstrom @carlbergstrom.com · 27/09/2026
In this case there's no reputational harm because everyone knows that Gemini is literally a bullshit machine—but if we allow these systems to serve as epistemic grounding, what are the remedies in cases where reputations are actually made or broken?
thebullshitmachines.com
Lesson 2: The Nature of Bullshit
This lesson explores philosopher Harry Frankfurt's notion of bullshit, and explains why LLMs are, by his definition, bullshit machines.
514029
Reposted by Matthew Chalmers
The Register @theregister.com · 27/09/2026
Big AI's content problem: Take the work, keep the money
theregister.com
Big AI's content problem: Take the work, keep the money
The more we learn about how AI does business, the more unfair it looks
12613
Reposted by Matthew Chalmers
Emily M. Bender @emilymbender.bsky.social · 26/09/2026
When Coxon was suddenly ALL OVER the media, it felt to many like that was because his story deserved that much interest i.e. that the media was doing its job informing the public of an important threat. The wall-to-wall media coverage wasn't organic though, it was orchestrated: >>
piratewires.com
618082
Reposted by Matthew Chalmers
Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 26/09/2026
A rich person can never be held accountable. Therefore, a rich person must never make a societal decision
899952571
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 26/09/2026
A massive Microsoft data centre was fined after a drone investigation revealed the use of illegally operating gas turbines -->> www.theguardian.com/environment/...
7357135
Matthew Chalmers @matthewchalmers.bsky.social · 26/09/2026
Augustine: “The good Christian should beware the mathematician and all those who make empty promises. The danger already exists that the mathematicians have made a covenant with the devil to darken the spirit and to confine man in the bonds of hell.” www.theguardian.com/technology/2...
theguardian.com
The dark art of maths and the AI revolution | Letter
Letter: Bev Littlewood reminds us that mathematics has been the prime driver of the rise in AI, in response to an editorial
011
Reposted by Matthew Chalmers
Carl T. Bergstrom @carlbergstrom.com · 26/09/2026
Does my Speak-n-Spell feel pain when I type O-U-C-H? The jury is out. Philosophers and neuroscientists say that's fucking stupid, but I called a guy from the Center for Pretending that Household Electronics Are Sentient and he said there's no way to know for sure.
5843441194
Reposted by Matthew Chalmers
Climate Defiance @climatedefiance.bsky.social · 24/09/2026
HUGE UPDATE: we halted Amazon's panel at Climate Week. We shut it down so hard that every guest and every speaker had to leave the premises. No one was able to return. Their whole event lasted less time than their warehouse workers' pee break.
25748256
Reposted by Matthew Chalmers
Justine Calma @justcalma.bsky.social · 24/09/2026
Talking about AI and fossil fuels, Huang says ‘in order to save you, they’ve got to hurt you first.' Huang’s extreme level of privilege is glaring (gift link) www.theverge.com/tech/1000140...
theverge.com
Jensen Huang talks about AI and climate change like a supervillain
It’s giving Thanos energy.
810028
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 23/09/2026
We did a deep dive into NVIDIA's wild claims here: www.euronews.com/2026/09/21/n... ketanjoshi.co/2026/09/16/n...
euronews.com
Nvidia chip use pollutes as much as a Russian coal giant, report says
The AI boom has driven US chipmaker Nvidia’s supply chain emissions up by 725% since 2020, according to a new report by environmental and civil society groups.
06023
Reposted by Matthew Chalmers
Research Professional News @resprofnews.bsky.social · 23/09/2026
Public support for research falls sharply when it uses AI, study finds. Scepticism over artificial intelligence in taxpayer-funded research “demonstrates fragility” of support within UK. www.researchprofessionalnews.com/rr-news-uk-i...
researchprofessionalnews.com
Public support for research falls sharply when it uses AI, study finds - Research Professional News
Scepticism over artificial intelligence in taxpayer-funded research “demonstrates fragility” of support within UK
2116
Reposted by Matthew Chalmers
Dr Abeba Birhane @abeba.blacksky.app · 22/09/2026
i struggle to make sense of a world dictated by the stupid
4628
Reposted by Matthew Chalmers
Ketan Joshi @ketanjoshi.co · 22/09/2026
The climate impacts of generative AI are going to be front and centre at COP31 - good!!! www.reuters.com/sustainabili...
reuters.com
https://www.reuters.com/sustainability/cop/turkey-launches-ai-initiative-ahead-cop31-climate-summit-2026-09-21/
05412
Reposted by Matthew Chalmers
Dr Abeba Birhane @abeba.blacksky.app · 21/09/2026
"In a leaked document, the Irish Presidency proposes to EU Member States that the interest of OpenAI, Anthropic, Google, Meta and SpaceX to make profits should take precedence over the fundamental right to data protection" noyb.eu/en/ai-eu-mem...
noyb.eu
AI: EU Member States plan “digital expropriation” of Europeans in the interest of AI companies
In a leaked document, the Irish Presidency proposes to EU Member States that the interest of AI companies to make profits should take precedence over the fundamental rights of Europeans.
19468293
Matthew Chalmers @matthewchalmers.bsky.social · 19/09/2026
Agreed. This is incredible.
000