Prof. of Computer Science, U. Glasgow.
UK
Joined June 2011
- Tweets 9,114
- Following 262
- Followers 375
- Likes 5,981
Matthew Chalmers retweeted
We should ask ourselves if the unintended consequences of our (and my) enthusiasm for “open access” has been the extreme profit margin (funded by the public purse) of commercial publishers - and if so what should we do now ?
Matthew Chalmers retweeted
All this to
-Undercut creative jobs
-Cover the internet in AI slop
-Introduce unreliable and biased AI agents into the classroom
-Automate the writing of marketing materials
-Create the corniest images and memes
-Further concentrate power and control in Silicon Valley
New Georgia Power load forecasts just dropped. Your jaw will, too.
DATA CENTERS: 31 GW
MFG'ING: 1.2 GW
OTHER: 1.2 GW
CLEAN TECH: 1 GW
"Since the filing of the 2023 IRP Update in October 2023:
The total economic development pipeline has expanded from approximately 17,000 MW to
approximately 36,500 MW through the mid-2030s;
Near-term 2028/2029 large loads have increased by approximately 9,400 MW; and
Customer commitments have expanded from 3,612 MW to 8,046 MW."
Matthew Chalmers retweeted
In the New York Times' lawsuit against OpenAI, NYT lawyers spent 150 hours combing through OAI's training data. OpenAI engineers *erased it* and admitted to erasing it. The NYT is now unable to get that data back. This should be a serious scandal.
Matthew Chalmers retweeted
Revenue sharing for search default and/or preferential treatment - dead. Google's tens of billions to Apple - dead (last I saw it's about 15% of Apple's profits). Reminder, this all needs to be argued and approved by Court and then will get appealed. Still far off. /3
Matthew Chalmers retweeted
Bam. There is it. US Department of Justice has filed - requesting divestiture of Chrome as a remedy for court's finding against Google. Android at risk, too. /1
Matthew Chalmers retweeted
We've just made it much easier to reuse the data on Our World in Data
Here is an overview of our new features: enhanced data downloads and the Chart Data API — by my colleagues @DanyX23, @MarcelGerber9, and @charliegiattino
ourworldindata.org/easier-to…
Matthew Chalmers retweeted
Letting chatbots run robots ends as badly as you'd expect go.theregister.com/feed/www.…
Matthew Chalmers retweeted
The Washington Post obtained details about a $45 million plan run by Elon Musk's advisors to target Black, Jewish, Muslim voters and others with "misdirection" ads - often containing falsehoods about her positions washingtonpost.com/politics/…
Matthew Chalmers retweeted
🚨NEW: The European Commission has just published the first draft of the Code of Practice for general-purpose #AI model providers.
digital-strategy.ec.europa.e…
Matthew Chalmers retweeted
The chicanery in Tech is so complete that even its so-called "legal compliance" measures break the law
(Reminder)
iccl.ie/tag/tcf/
Matthew Chalmers retweeted
Photoshop FOSS alternative GNU Image Manipulation Program 3.0 nearly here go.theregister.com/feed/www.…
Matthew Chalmers retweeted
Among the most important charts for thinking clearly about the 2024 election — and politics in general.
The hardest divide to cross isn't left vs. right. It's interested vs. uninterested. dataforprogress.org/insights…
Matthew Chalmers retweeted
normal guy: "do you have suggestions for how i can dress for my business casual office in arizona?"
me: "have you considered dressing like the king of spain?"
I think this might soon be part of our undergraduate HCI course.
piped.video/watch?v=7Slib2bb…
Matthew Chalmers retweeted
Meta was hit with a €798 million ($841 million) fine by European Union regulators by tying its Facebook Marketplace service to its sprawling social network, the US tech giant’s first ever penalty for EU antitrust violations: bloomberg.com/news/articles/…
Meta hit with first ever EU fine for an abuse of dominance - €797.72 million for tying Facebook Marketplace to its sprawling social media network @business
Interesting new ACM HEALTH paper, and system, but it seems like there is no control group or similar. How much of the effect is due to just being in the trial, and how much is due to the system use?
doi.org/10.1145/3696425
Matthew Chalmers retweeted
Kasper @khornbaek talked about the theory crisis in HCI*.
My view: AI is the cause. “LLM accessories” won't do it, HCI needs to renew its theories, methods & models to center on human-AI interaction. Prove that they work better than trial and error.
* tinyurl.com/2p9hu7fv