Sylvain Miermont @miermont@qoto.org

"After a June 27 Goldman Sachs report finally said that the emperor, AI, had no clothes in a language tech grifters understand, the bubble has been starting to pop. OpenAI, valued at $80 billion, revealed it might be bankrupt by the end of the year, because AI is valueless and makes no money, while studies showed that products that mentioned AI in their marketing did less well, we may be witnessing the end of what tech reporter Ed Zitron calls The Rot Economy. And today it is tech--in the US, Japan, S. Korea and Taiwan--that is driving the market collapse."

Great analysis from @VickyACAB on today's #stockmarket news.

https://all-cats-are-beautiful.ghost.io/lets-all-watch-the-tech-economy-die/

"3 Degrees More" https://link.springer.com/book/10.1007/978-3-031-58144-1
- Shows the consequences of a global warming of +3°C
- Demonstrates that we still have it in our hands to ensure that global warming is limited to at least +2°C
- Details the plan of what politically feasible, cost-effective measures should now be taken
- This book is open access, which means that you have free and unlimited access

Some fun with reaction diffusion and cell noise

People larping in the city park on a summer evening as an ice cream truck rolls past

What's amazing to me about this article/study is that after discovering mentions of "AI" make people less interested in products, their recommendation is "so you should not call attention to the AI in your marketing". The idea of simply *not doing* the thing that people don't want is apparently never considered. We cannot conceive of a way of constructing software other than "develop for the investors" https://xoxo.zone/@vwampage/112881383051537497

If computer hardware innovation frozes today we would still have decades before we exhaust all possible performance improvements on software.

The software industry needs to smart up. The free ride has been terribly wasteful and harmful to the planet. Update cycles needs to slow down and we need to relearn how to code with performance in mind.

"One crawler downloaded 73 TB of zipped HTML files in May 2024, with almost 10 TB in a single day. This cost us over $5,000 in bandwidth charges, and we had to block the crawler. We emailed this company, reporting a bug in their crawler, and we're working with them on reimbursing us for the costs."
https://about.readthedocs.com/blog/2024/07/ai-crawlers-abuse/

RIP @HackSpacemag 😔

https://www.raspberrypi.com/news/a-farewell-from-hackspace-magazine/

Seeing some #fieldtelephone stuff here again, which brought up an older thought again

Are there any DIY field telephone kits/PCBs/schematics out there, possibly open source hardware?

I know there are schematics for most field telephones out there but it doesn’t seem like parts would be easily available. I was thinking of something like a PCB you could just plug a POTS phone receiver into. Anyone ever seen something like that? Or do I have to keep it on my project list 😅

#Feldtelefon #oshw

I'm in love with the New York Public Library's book train, which shuttles materials from the stacks below the Stephen A. Schwarzman Building to researchers upstairs.

That's just one example of the co-evolution of books, libraries, trains, and transit...

📚🚋🧵

Hot take: public transit should not be run "for profit". The profit comes from everybody getting to their destination quickly, efficiently, and predictably.

For anyone who needs some relaxing distraction this afternoon, here is a large bubble in a viscous fluid bumping into a smaller one, and forcing it to distort before it engulfs it (taken at the Catalyst science centre in Widnes). I played with this for AGES... so much fun!

#bubbles #science

@KevinMarks Hmm, it's almost as if the people who organise their life around the relentless pursuit of resource acquisition at all costs are not the people whose control over capital provides the greatest societal benefit.

Reporting of the latest UBI test programme is ignoring the key finding, obvious to anyone who knows low income people, that their first impulse was to spread the wealth to others.
This thinking is so alien to economists and financial types that they barely discuss it. https://www.openresearchlab.org/findings/key-findings-spending

For all the talk of AI replacing everything, today proves the value of a solid technical support person that can leg it around the building getting things done. The least fashionable role in IT was the most important. The worker.
#Crowdstrike

I am kinda impressed with the 100% paper packaging, not even a plastic film on the display

I was supposed to get a call from a reporter to talk about space junk right after the goat sale, but he just emailed to say that Toronto is flooding and he just got off the phone with Weather Canada and has to write that up right now.

Which is so extremely 2024: "Sorry I can't do the interview about garbage that billionaires dropped on you from orbit, climate change is causing too many problems in my city right now"

I think reasonable people can disagree on the nuances of privacy engineering here...you can get really in the weeds of epsilons and trust boundaries.

Mozilla need to understand that that isn't what is happening here.

The fact is this is a new data gathering vector, it doesn't matter how privacy-preserving it is, it should be subject to *new, informed, and proactive* consent - rather than being automatically enrolled in an experiment.

Two updates to this thread.

Update 1: In this thread I complain Mozilla does not provide specific technical details about this feature. It turns out there *is* a document with the technical details, on Github:

https://github.com/mozilla/explainers/tree/main/ppa-experiment

It also explains (https://wiki.mozilla.org/Origin_Trials) which sites are participating in the feature.

I am linking this document because I believe the first five words do more to discredit what Mozilla is doing here than anything I could say:

"Mozilla is working with Meta"