Anthropic is different because they're committed to safety.

...I'm being told they are no longer committed to safety.

https://time.com/7380854/exclusive-anthropic-drops-flagship-safety-pledge/

The United Kingdom is escalating its censorship and mass surveillance. When Mullvad tried to criticise this with the TV ad “And Then?”, it was banned on British television.

https://media.mullvad.net/andthen/30s/Mullvad_AND_THEN_30s_US_ProRes422_Webmix_4K.mov

📊 Saiba Como Monitorar e Gerenciar Processos no Linux 🐧

Manter o controle dos processos é essencial para garantir o bom funcionamento do seu sistema. Aprenda como monitorar e gerenciar processos no Linux de forma eficiente com ferramentas práticas e dicas úteis!

👉 Leia mais no blog: https://nova.escolalinux.com.br/blog/saiba-como-monitorar-e-gerenciar-processos-no-linux?utm_source=dlvr.it&utm_medium=mastodon

#Linux #Administração #Processos #Tecnologia #OpenSource

Microsoft has pushed back against claims that multiple prompt injection and sandbox-related issues raised by a security engineer in its Copilot AI assistant constitute security vulnerabilities. The development highlights a growing divide between how vendors and researchers define risk in generative AI systems.

https://www.bleepingcomputer.com/news/security/are-copilot-prompt-injection-flaws-vulnerabilities-or-ai-limits/

In two recent public speaking engagements, I made a comparison between the obesity epidemic in various pockets of the globe and AI. Several people came up afterwards and said that was the best analogy they'd heard, so maybe it's worth repeating here. Also, I feel like this one will hold up a while.

What I said was some recent CDC figures show that something like 55 percent of the average American diet comes from processed or highly processed "foods," and the numbers are even higher for kids. In a similar way, so much of what AI produces is akin to highly processed data: Its myriad origins are murky at best, you often don't feel great after using it a while, and if you are exposed to it too much it might just freaking kill you.

1: "Well ChatGPT says otherwise. It contradicts what you said about security."
0: "ChatGPT is consuming popular material from 10 years ago. The threat landscape has changed."

1: "Well the language is clear it's saying the exact opposite of you."
0: "It has made a statement. But what is it's reasoning. What is the logical argument behind its suggestion?"

1: "I just feel more comfortable going with what it say."
0: "After we've worked together for five years? After you've seen the results of my work? You realize this isnt your field of expertise? You arent able to explain its reasoning. Why it said it. The history and context behind it."

Well this is going to be fun.

Being a leader in your field and releasing new research that is different from the current consensus now has to fight AI and LLMs.

AI / LLMs is bias at scale. Now it's entrenched status quo.

AI cannot innovate. AI can only average out. And when that average is bad. Or dangerous. And speaks with authority. Folks will double down and be hurt.

Do not stay at a Hyatt hotel. They're running fake "smoking detectors" specifically as a revenue-generating scheme and arbitrarily fining guests $500.

https://threadreaderapp.com/thread/1945959030851035223.html

After helping five friends in the last two months transition from #windows to #linux here are my thoughts:

1. Show case both #kde and #gnome before you pick it form them. 2/5 Decided to try KDE because it reminded them of Windows and the rest wanted Gnome because they where tired of windows UI
2. Do not fill their head with distros, go with one where it is easy to get support. I went with @fedora as it is the one I know best and like the resource for. So far only one has swapped (to Ubuntu)

Mankind is doomed

https://hackerone.com/reports/3250490