Notices by funnymonkey (funnymonkey@freeradical.zone)

I love it when defenders of AI claim that automated decision making will take human inaccuracies out of the decision making process.

pEoPLE hAvE b33N roNg aLL th3 TiME!!

AI provides the foundational tools for more people to be more wrong at a higher speed with less transparency at a significantly higher carbon footprint, using tools created by abusing and underpaying workers.

A 25 year old with no government experience has admin rights to top secret systems that control payments from the US government.

Co-presidents Musk and Trump are creating multiple single points of failure to allow sensitive and classified information to leak.

https://www.wired.com/story/elon-musk-associate-bfs-federal-payment-system/

EDIT: to specify the amount of money involved

@tek Anytime a site confesses they use Admiral, they can fuck right off.

This is exactly what Microsoft did!

No one wants this crap, they need to justify their investment (and get access to more data) so they increase their userbase by forcing it on people as a "feature"

The tech industry continues to fail -- utterly, miserably -- at consent.

https://techcrunch.com/2025/01/15/google-raises-the-price-of-workspace-plans-making-its-ai-features-included-for-free/

Can anyone point me to any good, reasonably current writing on the relationship between industry and academia in AI research?

Reminder (largely to myself, but to anyone else who needs to hear it):

Pace yourself.

The despicable, the disgusting, the hateful, the criminal, will be coming at us from all corners.

It is, and will be, exhausting. Choose your battles. Choose your friends wisely. Save energy. Love those who deserve it.

But yeah - there is a difference from knowing "it will be bad" to seeing the specific shapes and forms bad will take.

Pace yourself.

Striking down net neutrality isn't a "blow to Democrats", as the Washington Post headline reads.

It's a blow to anyone who wants the internet to stay accessible.

Killing net neutrality is a gift to people and companies that want to limit, restrict, throttle, and otherwise shape our access.

It's not a Democrat versus Republican thing - it's about maintaining access to core services.

Apple takes privacy very seriously lol sure.

And 95 million to Apple in a year isn't much, and spread over ten years is nothing.

Violating our privacy, getting caught, and paying fines is the cost of doing business.

https://arstechnica.com/tech-policy/2025/01/apple-agrees-to-pay-95m-delete-private-conversations-siri-recorded/

This and more.

What if Musk determines that a Tesla owner might be a political opponent?

Or a business rival?

Or might be saying something interesting?

Unethical people and companies do unethical things.

https://www.404media.co/elon-musk-uses-cybertruck-explosion-to-show-tesla-can-remotely-unlock-and-monitor-vehicles/

Why, oh why, do sites do this?

When we evaluate the customer facing security of sites, "features" like this need to be explicitly flagged as undermining basic security practice -- in this case, underminig 2FA.

Google wants to thank you for sharing all of your output with them.

And, by extension, law enforcement, governments, and legal requests.

https://www.theverge.com/2024/12/12/24319697/gemini-ai-summarize-google-drive-folders

Does anyone know how many AI spam bug reports come through the bug bounty platforms (HackerOne, etc)?

"We're going to create nuclear waste and ruin slightly less water as we power our tech that doesn't work" isn't the flex you think it is.

Looking for sites that support local bookstores - I have come across bookshop.org -

Can anyone recommend other sites/services?

(I'm thinking specifically of services that people can use when they don't have easy access to a locally owned bookstore)

Thanks for any recs!

@tek is the absolute best.

There are Republican senators cosplaying Very Serious People as they hem and haw about the string of mean spirited and highly incompetent people Trump is putting forward to run multiple federal agencies.

My prediction (and I hope I'm wrong - I really really do): these senators will all fall into line, kiss the ring, and put these incompetent nominees into positions they are uniquely and dangerously unqualified to fill.

Note: when you read an article where the writer says that a feature can be used for "fingerprinting" or a "device profile", and then says that the fingerprint/profile can be used for targeted ads:

That journalist (or their editor) has an incomplete understanding. Targeted ads are a symptom. The main problem is that if you can be targeted with an ad, you can also be identified from within the crowd.

Adtech is a symptom of more pervasive privacy problems.

For no apparent reason aside from the obvious, I'm updating a personal privacy and security guide I wrote back in 2015-2016.

It's not done yet, but it's getting there. Hopefully, I'll have the details set tomorrow.

https://github.com/billfitzgerald/personal-privacy-security

The focus of the guide is to collect up practical and effective steps - it's not designed as a comprehensive guide to going dark, but as a menu of options people can choose from over time to minimize risk.

The best privacy and security changes we can make are the ones we will actually make and stick to.

2/x

Back in 2016/2017 I co-wrote a guide to personal privacy and security. I've been meaning to update it for a while, and made some minor edits in June 2024, but I finally took the plunge this week and made some additional updates.

This doc is still a work in progress, but I was able to get it adequately clean and updated to the point where it's usable in its current form. I'll keep adding to it and tweaking it, but IMO it's clean enough to be useful.

1/x