Top Stories Daily

Selfish AI | GarfieldTech

garfieldtech.com · Feb 02

This will be a bit more ranty than my usual articles. Fair warning. But I need to put this out there.

Shared by @FaithfullJohn and 138 others.

LillyLyle/Count Melancholia (@LillyHerself) · Feb 03

🔁 @Crell:

I may regret this at some point, but I felt the need to put down in writing how I feel about this moment in the tech industry.

It is not kind. You may well be insulted by it. If you are... then you really should question yourself.

https://www.garfieldtech.com/blog/selfish-ai

#AI #LLM #Programming

Hide

Getting off US tech: a guide

disconnect.blog · Feb 03

I’m in the process of dropping US tech services. Here’s how I did it, and options you should consider.

Shared by @trendytoots and 29 others.

brightside (@brightside) · Feb 03

🔁 @steveroyle:

If you're interested in divesting yourself of US-based tech and reducing your digital footprint, there's plenty of guides to help (like these two).

https://disconnect.blog/getting-off-us-tech-a-guide/

https://www.optoutproject.net/tags/cleanse/

Hide

Copilot everywhere? Not for long. Microsoft dialing it back on Windows 11

windowscentral.com · Jan 30

People familiar with Microsoft's plans say that the company moving to streamline or remove certain Copilot integrations across in-box apps like Notepad and Paint in 2026, after pushback from users.

Shared by @joelpomales and 56 others.

Darrell Hilliker 👨‍🦯♾️📡 (@darrell73) · Feb 03

🔁 @ai6yr:

Windows Central: You won: Microsoft is walking back Windows 11’s AI overload — scaling down Copilot and rethinking Recall in a major shift

https://www.windowscentral.com/microsoft/windows-11/microsoft-is-reevaluating-its-ai-efforts-on-windows-11-plans-to-reduce-copilot-integrations-and-evolve-recall

#windows #AI #copilot

Hide

He lost an eye in the George Floyd protests. Now he represents the ward where Renée Good was killed.

motherjones.com · Jan 30

Meet Minneapolis city council member Soren Stevenson.

Shared by @RhinosWorryMe and 17 others.

Lauralee Dukeshire (@LauraleeDukeshire) · Feb 02

🔁 @nullagent:

Incredible interview with Minneapolis city councilman Soren Stevenson who lost an eye to a "rubber" bullet shot at him by MPD during the Goerge Floyd protests.

"Truly, I see this as a moment when we are deciding right here in Minneapolis—in this country—are we going to have a democracy going forward, or are we not? The stakes are no greater and no smaller than that."

https://www.motherjones.com/politics/2026/01/minneapolis-renee-good-alex-pretti-soren-stevenson-council-interview/

#minnesota #SorenStevenson #GeorgeFloyd #minneapolis

Hide

A century of hair samples proves leaded gas ban worked

arstechnica.com · Feb 02

“We should not forget the lessons of history. And the lesson is those regulations have been very important.”...

Shared by @rpsu and 24 others.

Scott Francis (@darkuncle) · Feb 03

🔁 @ScienceDesk:

In the 1970s, the U.S. Environmental Protection Agency (EPA) cracked down on lead-based products including gasoline because of their toxic effects on human health. Now, scientists at the University of Utah have released the findings of a study looking at 100 years' worth of human hair samples, and found that the regulatory action worked. Here's more from @arstechnica.

https://flip.it/sslA6i

#Science #Environment #EnvironmentalProtectionAgency

Hide

Docusign’s CEO on the dangers of trusting AI to read, and write, your contracts

theverge.com · Feb 02

Docusign’s Allan Thygesen says his company’s pivot to AI is a necessity in the world of contract management.

Shared by @Techmeme and 7 others.

SuffolkLITLab (@SuffolkLITLab) · Feb 02

TL;DR: Docusign CEO Allan Thygesen discusses the risks of relying on AI for contract management, highlighting the complexities involved in automating legal processes. With the company employing 7,000 people, he sheds light on the crucial human oversight needed in this technology-driven arena. https://www.theverge.com/podcast/871205/docusign-ceo-allan-thygesen-ai-contracts-e-signature-interview #law #tech #legaltech ⚖️ 🤖 #autosum

Hide

Here's how Epstein broke the internet

garbageday.email · Feb 03

His meeting with the founder of 4chan and his quest to profit off the end of democracy

Shared by @Compassionatecrab and 12 others.

SilenceisGolden (@silentLurker) · Feb 03

🔁 @Nshrubs:

Here's how Epstein broke the internet
His meeting with the founder of 4chan and his quest to profit off the end of democracy
https://www.garbageday.email/p/here-s-how-epstein-broke-the-internet

Hide

WhatsApp Encryption, a Lawsuit, and a Lot of Noise

blog.cryptographyengineering.com · Feb 02

It’s not every day that we see mainstream media get excited about encryption apps! For that reason, the past several days have been fascinating, since we’ve been given not one but sever…

Shared by @stereo and 10 others.

ralf tauscher @ fosdem (@stereo) · Feb 03

🔁 @david_chisnall:

A few things about this article on the claims WhatsApp's E2EE is fake:

The most important thing to keep in mind here is that Meta’s encryption happens on the client application, the one you run on your phone. If the claims in this lawsuit are true, then Meta would have to alter the WhatsApp application so that plaintext (unencrypted) data would be uploaded from your app’s message database to some infrastructure at Meta, or else the keys would

They would not have to be sent as plaintext. They would be sent encrypted with some key that Meta owns.

The other simple way of doing this is to intentionally weaken the key generation function. If you are generating a key with 128 bits of entropy, you can start with a 96-bit secret shared between the server and the client. Then you add 32 bits of random number and feed the result through some cryptographic hash function. You end up with a key that, to an outside observer, is one of 2128 possible values. But to the server operator, it is one of 232 possible values and so takes a few seconds of CPU time to recover. And there are ways of doing this that look a lot like an honest mistake. Telegram has a thing in their protocol for the server to provide part of the entropy seed, which they claim is intended to provide additional defence for devices with weak entropy sources and critics argue is for exactly this kind of attack. Putting the same kind of entropy seed into an out-of-band request to another Meta service that WhatsApp uses would be quite feasible.

And there are ways of hiding this kind of thing from various forms of forensic analysis, such as dynamically patching the random number generator from another thread when some trigger is applied. It's not impossible to find, it's just really hard, and someone would have to be looking for the right thing.

There's also WhatsApp Web. This runs something in the web browser that talks to your phone and uses end-to-end encryption with a key exchanged by a QR code shown in the browser. Signal does not implement a feature like this because designing it in such a way that it's verifiably (by the user) secure is almost impossible.

David Chisnall (*Now with 50% more sarcasm!*) (@david_chisnall) · Feb 03

A few things about this article on the claims WhatsApp's E2EE is fake:

The most important thing to keep in mind here is that Meta’s encryption happens on the client application, the one you run on your phone. If the claims in this lawsuit are true, then Meta would have to alter the WhatsApp application so that plaintext (unencrypted) data would be uploaded from your app’s message database to some infrastructure at Meta, or else the keys would

They would not have to be sent as plaintext. They would be sent encrypted with some key that Meta owns.

The other simple way of doing this is to intentionally weaken the key generation function. If you are generating a key with 128 bits of entropy, you can start with a 96-bit secret shared between the server and the client. Then you add 32 bits of random number and feed the result through some cryptographic hash function. You end up with a key that, to an outside observer, is one of 2128 possible values. But to the server operator, it is one of 232 possible values and so takes a few seconds of CPU time to recover. And there are ways of doing this that look a lot like an honest mistake. Telegram has a thing in their protocol for the server to provide part of the entropy seed, which they claim is intended to provide additional defence for devices with weak entropy sources and critics argue is for exactly this kind of attack. Putting the same kind of entropy seed into an out-of-band request to another Meta service that WhatsApp uses would be quite feasible.

And there are ways of hiding this kind of thing from various forms of forensic analysis, such as dynamically patching the random number generator from another thread when some trigger is applied. It's not impossible to find, it's just really hard, and someone would have to be looking for the right thing.

There's also WhatsApp Web. This runs something in the web browser that talks to your phone and uses end-to-end encryption with a key exchanged by a QR code shown in the browser. Signal does not implement a feature like this because designing it in such a way that it's verifiably (by the user) secure is almost impossible.