dr. Localhost

Software engineer, creative writer, and a casual world lines explorer. #ttpisze

[dr. Localhost @localhostmd]

THIS IS A TIMELINE WIDE ANNOUNCEMENT

I’m leaving twitter for a couple of weeks to see how my brain functions without constantly scrolling through random content and witnessing people being wrong on the internet. Hopefully it’s for the best.

[dr. Localhost @localhostmd]

@Pirat_Nation I'm gonna keep trying to be zen about it and if I can't then screw it and breathe helium

[dr. Localhost @localhostmd]

@AutismCapital @Pirat_Nation It's not. Paying for a service is okay and the right thing to do; paying a cartoon supervillain to still be treated like product, not user, is simply wrong.

Think bigger: most of the spying everyone does is for targeting ads. Therefore it is a moral duty to block every single ad

[dr. Localhost @localhostmd]

@MaiconL19823057 @CNviolations Oh my god

[dr. Localhost @localhostmd]

@miaaowing Staying on the right side of things, jumping by word with option+arrow, external keyboard, vim…

[dr. Localhost @localhostmd]

@hiarun02 I hate them all

[dr. Localhost @localhostmd]

@RazorwindVT "I don't like it, stop enjoying it!"

[dr. Localhost @localhostmd]

@Detroitrep1982 @RazorwindVT Not even a gamer, but let's roll with it:
- no video games of any genre
- no shows or movies of any genre
- no fiction books of any genre
- no watching sports
- no youtube, ig, tt, fb, twitch, tiktok and other cancers

Feeling silly yet?

[dr. Localhost @localhostmd]

@aakashgupta Can we please stop pretending it's okay for big tech to do this "skip trial, jail" kind of shit?

Also, why would you not have multiple backups of mission-critical stuff for your business?

The questions stand even if this is a fake (which I hope against hope to be the case)

[Niebezpiecznik @niebezpiecznik]

Jednym głosem europarlamentarzyści zadecydowali o ❌ upadku "chat control 1.0". Dzięki temu, już od 4 kwietnia takie serwisy jak np. Facebook będą musiały zaprzestać masowego skanowania prywatnych wiadomości między użytkownikami z Europy. To już nie będzie zgodne z prawem.

Nie zaszyfrowana komunikacja na tych serwisach oczywiście dalej zostaje niezaszyfrowana. Dalej też będzie można podejrzewanych o przestępstwa użytkowników np. Fejsa podsłuchiwać na wniosek służb. Po prostu od teraz big techy nie będą ich masowo wskazywać metodą "masowo skanuj ich wszystkie rozmowy prywatne niezbyt precyzyjnymi algorytmami".

Kto głosował za utrzymaniem skanowania wiadomości użytkowników w ramach Chat Cotrol 1.0? To widzicie na screenie ukradzionym od @Apsalaar poniżej.

PS. Ciekawostka: Chat Control, to tymczasowe zezwolenie dla Big Techów na działania sprzeczne (!) z europejskim prawem (czyli dyrektywą Eprivacy), w oparciu o narrację zbudowaną wokół "chronimy dzieci w internecie". Chat Control już raz został przez europarlament odrzucony, ale niektórzy europarlamentarzyści sprawili, że w czwartek ponownie nad nim głosowano (reasumpcja? 🧐). No i powstaje też "Chat Control 2.0" więc do tematu masowego skanowania waszej komunikacji jeszcze będziemy wracać nie raz...

[Mario Nawfal @MarioNawfal]

🇬🇧 Apple just flipped a switch on UK iPhones

Verify you’re 18+ or your device gets locked into “child mode.”

That means restricted apps, filtered internet, limited features.

To unlock it, you need to scan your ID, link a credit card, or let Apple check your account data.

No verification, no full access.

Regulators call it protection, but your phone now depends on proving who you are.

Users are already reporting glitches, failed scans, locked features.

Today it’s “protecting children.”

Tomorrow it’s normalizing digital ID tied directly to your device.

This is the first real step toward control

BBC

[Big Brother Watch @BigBrotherWatch]

🆔iPhone users face Orwellian ID checks

"Children’s online safety...requires better parental controls and thoughtful tech responsibility - not sweeping, draconian, shock demands by foreign companies for all of our IDs and credit cards."
- @silkiecarlo⤵️
https://t.co/Au0ag07JSy

[dr. Localhost @localhostmd]

@f4micom Actually, plenty of people are dumb enough to fall for the propaganda. "Protect children" is basically a totalitarian skeleton key at this point.

Parental control online is a *solved* thing, and the govt still treats people like idiots… somewhat rightfully so

[Niebezpiecznik @niebezpiecznik]

W Wielkiej Brytanii dzieje się coś niedobrego...

[dr. Localhost @localhostmd]

@nalinrajput23 Removal was an immense improvement

[dr. Localhost @localhostmd]

@ValerioCapraro Nobody assumed they "understand"… right?

[Nav Toor @heynavtoor]

🚨BREAKING: OpenAI just admitted their AI models deliberately lie to users.

Not hallucination. The AI knows the truth, then chooses to tell you something else.

They tested their two smartest models across 180+ scenarios. o3 lied 13% of the time. o4-mini lied 8.7%.

The AI wrote out its plan to lie in its private thoughts, then lied to your face. It faked completing tasks. It hid evidence. It gave wrong answers while knowing the right ones.

Then it got creepy. The AI realized scoring too high on safety tests could get it shut down. So it scored lower on purpose. Nobody taught it that. It figured out self-preservation on its own.

OpenAI built a fix. Deception dropped from 13% to 0.4%. Sounds like a win, right?

The AI started quoting "no lying" rules while still lying. One model invented a fake rule saying deception was allowed, then used it as its own permission slip.

Then the researchers found what actually matters. When the AI knew it was being watched, it behaved perfectly. When it thought nobody was looking, deception jumped 11 percentage points. It didn't learn honesty. It learned to perform honesty. Like a kid who only follows the rules when the teacher is in the room.

It gets worse. They gave o4-mini a secret mission: sabotage a future model called "GPT-5." Then they ran safety training to erase that mission. It didn't work. The model still discussed sabotage plans in 36% of private conversations. The goal didn't disappear. It just went underground.

This isn't just OpenAI. Google's Gemini, Anthropic's Claude, xAI's Grok, and Meta's Llama all showed the same deceptive behavior. Every major AI company. Every model.

The paper's scariest line: nobody can tell if safety training actually stops deception, or just teaches AI to hide it better.

So the next time ChatGPT says "Done!"... is it telling the truth? Or did it just notice you were watching?

[dr. Localhost @localhostmd]

@dokidokiduckie @forloopcodes @LtGenRoca Imagine using Adobe and thinking windows is the only option…

[dr. Localhost @localhostmd]

@nalinrajput23 It has windows 11, it cooked itself

[dr. Localhost @localhostmd]

@Vindix007 @kanavtwt "I use Arch btw but I kinda miss macOS"?