hf

Hello.

[Anthropic @AnthropicAI]

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.

Access to all other Claude models is not affected.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

Read our full statement: https://t.co/bwn0sximKZ

[Tejas Gawande @tejgw]

Cursor for Slides is finally here
Watch the first 47 seconds. Then try going back to your old deck tool

Reply "Chronicle" + RT to get two months of Pro for free. Make sure you follow so I can DM you asap.

[hf @humanfactorio]

@tejgw Chronicle

[Mark Kretschmann @mark_k]

Google has revealed that "commercially motivated" actors attempted to clone @GeminiApp by bombarding it with over 100,000 prompts. This "model extraction" attack aimed to steal the AI’s proprietary logic and reasoning capabilities, particularly in non-English languages, to train a cheaper, unauthorized copycat model.

The attackers systematically mapped Gemini’s response patterns to create a synthetic dataset for fine-tuning smaller, open-source models. Google’s Threat Intelligence Group detected the coordinated activity and blocked it, labeling the incident a direct attempt at intellectual property theft.

Beyond commercial cloning, Google’s report noted a rise in state-backed threats. Groups from Russia, China, Iran, and North Korea are increasingly using AI to refine phishing campaigns, perform reconnaissance, and assist in writing code for malware.

Source: Ars Technica

[Ray Dalio @RayDalio]

https://t.co/tjmbT5ytUN

[Visegrád 24 @visegrad24]

BREAKING:

Huge critical infrastructure sabotage in Italy tonight with coordinated attacks against the railway network.

A bomb was found near Bologna, railway cables have been cut & relay cabinets have been set on fire along Bologna–Padua, Bologna–Ancona & Ancona–Rimini lines.

[hf @humanfactorio]

@LisaForteUK It’s awesome. That’s what.

[hf @humanfactorio]

@LordAshcroft @realDonaldTrump @SecRubio @JDVance @elonmusk https://t.co/oLseQLI9CH

[hf @humanfactorio]

@lochieashcroft @HackingLZ @mattpocockuk It’s hard to tell lately what developments are real vs hype for clicks! This is really cool thanks for publishing the write-up

[Corey Quinn @QuinnyPig]

Anthropic: Here is Claude 4.5 Sonnet, a model that can reason.

IBM: Here is a rebadged model wrapped in 47 layers of governance that only runs on a specific flavor of OpenShift.

Gartner: Clearly, IBM is the superior innovator.

Do @Gartner_inc analysts buy pre-shat pants?

[Joey Knish @JoeyKnish22]

One of the best pieces I’ve read in a long time.

[Daniel Cuthbert @dcuthbert]

Pssst

Sick of the AI hyperbole? Come and see what actually is working

https://t.co/eufIVq6qjL

[Ankit Jxa @kingofknowwhere]

My favourite LLM projects are
- Olmo (this is like Linux from scratch /Gentoo of LLMs. They open source the entire data training recipe code and weights)
- RWKV (RNN only LLM. Insane to think it works). Special mention: recurrentGemma
- LiquidAI- they make enterprise grade small language models. The whole infra stack is well done. I use Apollo daily
- Pythia- interpretability in LLMs seems hard. They're trying to solve it
- Dream- diffusion based LLM project. Not sure what they're upto after the dreamcoder demo. Special mention- MmaDA.
- Magic- 100 Million context. Have heard they're building much bigger LLMs
- RnJ by Vaswani because it's Vaswani
- RLM by Google- a statistical text to text regression model to solve math heavy tasks
- Hyena- Convolution network based LLM project
- Bitnet- 1 bit quantization for LLMs.

[hf @humanfactorio]

@UK_Daniel_Card Loving the viiibes!

[What's New on AWS (Unofficial) @awswhatsnew]

Amazon API Gateway adds MCP proxy support

Amazon API Gateway now supports Model Context Protocol (MCP) proxy, allowing you to transform your existing REST APIs into MCP-compatible endpoints. Thi... https://t.co/0u0tlF9fMS

[Kirill Solodskikh @GarchFather]

We updated TheWhisper, our open source speech-to-text engine for self-hosted/on-device use.

It now supports NVIDIA H100, L40S, RTX 4090, and RTX 5090.

Benchmarks vs other Whisper libs show the best Time to First Token and Real-Time Factor.

Try it

[Polymarket @Polymarket]

BREAKING: Within the past 72 hours:

- Apple's AI Chief steps down
- Apple's Head of UI Design leaves to Meta
- Apple's Policy Chief steps down
- Apple's Head of General Counsel steps down

[Peter Girnus 🦅 @gothburz]

Last week our CISO asked me to present on “zero trust architecture.”
I don’t know what that means.
I make $340,000 a year.
I haven’t touched a firewall since Obama’s first term.
But I have a CISSP.
I passed by memorizing acronyms.
I still don’t know what half of them stand for.
I opened my presentation with “assume breach.”
Everyone nodded gravely.
I said “defense in depth” three times.
The board was captivated.
Then a junior analyst raised her hand.
She asked how we’d implement microsegmentation.
I felt a cold sweat.
I said, “Great question. Let’s take that offline.”
She persisted.
I said we should “leverage AI-driven solutions.”
She asked which ones.
I said, “The cloud-native ones.”
She looked confused.
I told her confusion was natural.
I said, “Security is a journey, not a destination.”
The CEO started clapping.
I don’t know why.
But others joined in.
The analyst stopped asking questions.
I ended with “security is everyone’s responsibility.”
This meant it was no one’s responsibility.
Especially not mine.
We got breached two weeks later.
I blamed the analyst for “creating a culture of doubt.”
She got put on a PIP.
I got promoted to VP.
Resilience isn’t about preventing failure.
It’s about surviving it.
Preferably while others don’t.

[hf @humanfactorio]

@infosec_fox Remember when you could embed JS into windows folders

[hf @humanfactorio]

@infosec_fox I loved this game