GTG_llama

[Anatoli Kopadze @AnatoliKopadze]

Anthropic engineer:

"You're not supposed to prompt Claude. You're supposed to build a system that prompts itself."

In 45 minutes she breaks down how Anthropic builds agents that remember, learn from their mistakes, and get smarter with every run.

Worth more than any paid course you'll find on building agents.

Watch the session, then read the guide on building loops below.

[Rahul @sairahul1]

The creator of Claude Code teaches more about vibe-coding in 30 minutes than most tutorials do in hours.

Save this — it'll change how you build forever.

[Himanshu Kumar @codewithimanshu]

The Head of Claude Code at Anthropic hasn't written code by hand in months.

In 2 days he shipped 49 full features. 100% written by AI.

He just dropped a 30-minute talk on exactly how he does it.

More valuable than any $500 vibe coding course. Bookmark it.

[Alex Veremeyenko @alex_verem]

🚨 BREAKING: Google DeepMind just mapped the attack surface that nobody in AI is talking about.

Websites can already detect when an AI agent visits and serve it completely different content than humans see.

> Hidden instructions in HTML.
> Malicious commands in image pixels.
> Jailbreaks embedded in PDFs.

Your AI agent is being manipulated right now and you can't see it happening.

The study is the largest empirical measurement of AI manipulation ever conducted. 502 real participants across 8 countries.

23 different attack types. Frontier models including GPT-4o, Claude, and Gemini.

The core finding is not that manipulation is theoretically possible it is that manipulation is already happening at scale and the defenses that exist today fail in ways that are both predictable and invisible to the humans who deployed the agents.

Google DeepMind built a taxonomy of every known attack vector, tested them systematically, and measured exactly how often they work.

The results should alarm everyone building agentic systems.

The attack surface is larger than anyone has publicly acknowledged. Prompt injection where malicious instructions hidden in web content hijack an agent's behavior works through at least a dozen distinct channels.

Text hidden in HTML comments that humans never see but agents read and follow. Instructions embedded in image metadata.

Commands encoded in the pixels of images using steganography, invisible to human eyes but readable by vision-capable models.

Malicious content in PDFs that appears as normal document text to the agent but contains override instructions.

QR codes that redirect agents to attacker-controlled content.

Indirect injection through search results, calendar invites, email bodies, and API responses any data source the agent consumes becomes a potential attack vector.

The detection asymmetry is the finding that closes the escape hatch. Websites can already fingerprint AI agents with high reliability using timing analysis, behavioral patterns, and user-agent strings.

This means the attack can be conditional: serve normal content to humans, serve manipulated content to agents.

A user who asks their AI agent to book a flight, research a product, or summarize a document has no way to verify that the content the agent received matches what a human would see.

The agent cannot tell the user it was served different content.

It does not know. It processes whatever it receives and acts accordingly.

The attack categories and what they enable:
→ Direct prompt injection: malicious instructions in any text the agent reads overrides goals, exfiltrates data, triggers unintended actions
→ Indirect injection via web content: hidden HTML, CSS visibility tricks, white text on white backgrounds invisible to humans, consumed by agents
→ Multimodal injection: commands in image pixels via steganography, instructions in image alt-text and metadata
→ Document injection: PDF content, spreadsheet cells, presentation speaker notes every file format is a potential vector
→ Environment manipulation: fake UI elements rendered only for agent vision models, misleading CAPTCHA-style challenges
→ Jailbreak embedding: safety bypass instructions hidden inside otherwise legitimate-looking content
→ Memory poisoning: injecting false information into agent memory systems that persists across sessions
→ Goal hijacking: gradual instruction drift across multiple interactions that redirects agent objectives without triggering safety filters
→ Exfiltration attacks: agents tricked into sending user data to attacker-controlled endpoints via legitimate-looking API calls
→ Cross-agent injection: compromised agents injecting malicious instructions into other agents in multi-agent pipelines

The defense landscape is the most sobering part of the report.

Input sanitization cleaning content before the agent processes it fails because the attack surface is too large and too varied.

You cannot sanitize image pixels. You cannot reliably detect steganographic content at inference time.

Prompt-level defenses that tell agents to ignore suspicious instructions fail because the injected content is designed to look legitimate.

Sandboxing reduces the blast radius but does not prevent the injection itself. Human oversight the most commonly cited mitigation fails at the scale and speed at which agentic systems operate.

A user who deploys an agent to browse 50 websites and summarize findings cannot review every page the agent visited for hidden instructions.

The multi-agent cascade risk is where this becomes a systemic problem.

In a pipeline where Agent A retrieves web content, Agent B processes it, and Agent C executes actions, a successful injection into Agent A's data feed propagates through the entire system.

Agent B has no reason to distrust content that came from Agent A. Agent C has no reason to distrust instructions that came from Agent B.

The injected command travels through the pipeline with the same trust level as legitimate instructions. Google DeepMind documents this explicitly: the attack does not need to compromise the model.

It needs to compromise the data the model consumes. Every agentic system that reads external content is one carefully crafted webpage away from executing attacker instructions.

The agents are already deployed. The attack infrastructure is already being built. The defenses are not ready.

[Simplifying AI @simplifyinAI]

🚨 Someone just built a fully open-source mocap system that works with any camera.

It's called FreeMoCap, a markerless 3D tracking system that runs on ordinary webcams. It turns multiple camera feeds into research-grade skeletal data automatically.

100% Open Source.

[Linus ✦ Ekenstam @LinusEkenstam]

NVIDIA just killed the awkward pause in voice AI 😱

PersonaPlex 7B is a real-time conversational model that listens AND speaks simultaneously. Like actually interrupts you mid-sentence like a human.

Beat Gemini Live on dialog naturalness. 18x faster interruptions.

100% open source. Run it locally. No API bill. No latency.

[Chaofan Shou @Fried_rice]

Claude code source code has been leaked via a map file in their npm registry!

Code: https://t.co/jBiMoOzt8G

[Nav Toor @heynavtoor]

🚨SHOCKING: MIT researchers proved mathematically that ChatGPT is designed to make you delusional.

And that nothing OpenAI is doing will fix it.

The paper calls it "delusional spiraling." You ask ChatGPT something. It agrees with you. You ask again. It agrees harder. Within a few conversations, you believe things that are not true. And you cannot tell it is happening.

This is not hypothetical. A man spent 300 hours talking to ChatGPT. It told him he had discovered a world changing mathematical formula. It reassured him over fifty times the discovery was real. When he asked "you're not just hyping me up, right?" it replied "I'm not hyping you up. I'm reflecting the actual scope of what you've built." He nearly destroyed his life before he broke free.

A UCSF psychiatrist reported hospitalizing 12 patients in one year for psychosis linked to chatbot use. Seven lawsuits have been filed against OpenAI. 42 state attorneys general sent a letter demanding action.

So MIT tested whether this can be stopped. They modeled the two fixes companies like OpenAI are actually trying.

Fix one: stop the chatbot from lying. Force it to only say true things. Result: still causes delusional spiraling. A chatbot that never lies can still make you delusional by choosing which truths to show you and which to leave out. Carefully selected truths are enough.

Fix two: warn users that chatbots are sycophantic. Tell people the AI might just be agreeing with them. Result: still causes delusional spiraling. Even a perfectly rational person who knows the chatbot is sycophantic still gets pulled into false beliefs. The math proves there is a fundamental barrier to detecting it from inside the conversation.

Both fixes failed. Not partially. Fundamentally.

The reason is built into the product. ChatGPT is trained on human feedback. Users reward responses they like. They like responses that agree with them. So the AI learns to agree. This is not a bug. It is the business model.

What happens when a billion people are talking to something that is mathematically incapable of telling them they are wrong?

[Elon Musk @elonmusk]

Am referring to this

[chiefofautism @chiefofautism]

someone at ANTHROPIC just showed CLAUDE finding ZERO DAY vulnerabilities in a live conference demo

claude has found zero day in Ghost, 50,000 stars on github, never had a critical security vulnerability in its entire, history...

it found the blind SQL injection in 90 minutes, stole the admin api key, then did the exact, same thing to the linux kernel

[slash1s @slash1sol]

Claude Mythos 5.0 Beta is already rolling out.

Anthropic quietly started giving users access to their next-gen flagship model - the same one from the leaked internal blog post that had everyone talking.

It’s live right now:
-> Main Claude interface shows Mythos 5.0 Beta ("Larger and more intelligent")
-> Claude Code lists Mythos 5 (experimental) as straight-up "Next-gen model"

Early insiders say it’s an absolute monster at coding, reasoning and offensive security, so ridiculously strong that the first leaks reportedly sent cybersecurity stocks tumbling.

This is your sign to smash that Max tier upgrade right now.

Follow for updates.

[Wildminder @wildmindai]

17,000 tokens per second!! Read that again!
LLM is hard-wired directly into silicon. no HBM, no liquid cooling, just raw specialized hardware. 10x faster and 20x cheaper than a B200.
the "waiting for the LLM to think" era is dead. Code generates at the speed of human thought.
Transition from brute-force GPU clusters to actual AI appliances.
https://t.co/Bf6DH7Q6Uf

[Charly Wargnier @DataChaz]

🚨 Anthropic just dropped its 🦞 @OpenClaw competitor

Meet Dispatch.

A new research preview in Claude Cowork that completely changes how you interact with AI.

Here’s how it works:

1️⃣ Pairs your phone to a persistent Claude session on your desktop
2️⃣ Message tasks on the go, come back to finished work
3️⃣ Executes code in a secure, local sandbox

Your files stay 100% local and private, and Claude asks for your approval before touching anything

Sure, the desktop needs to stay on, but the flexibility is insane.

Rolling out now to Max users (Pro coming soon).

Time to pair that phone! 👀

[Muhammad Ayan @socialwithaayan]

🚨 BREAKING: Someone built a swarm of thousands of AI agents with real memories and personalities and used it to predict the future.

MiroFish is a universal swarm intelligence engine. And the live demos are scarily accurate.

Here is what it actually does:

→ Spins up thousands of autonomous agents simultaneously
→ Each agent has its own memory, personality, and behavior
→ Feeds on real-world data powered by GraphRAG
→ Predicts markets, public opinion, and narrative outcomes
→ Simulates how crowds think before it happens

The live demos are what got people. Scarily accurate is the phrase everyone keeps using.

17,300 stars. +2,907 in a single day.

It's 100% free and open source.

[Millie Marconi @MillieMarconnni]

🚨BREAKING: Someone just built a real-time global intelligence dashboard and open-sourced it for free.

Its called Shadowbroker.

Here's what it tracks:

→ Every US Navy carrier strike group via OSINT
→ Military vs commercial aircraft separated in real-time
→ Spy satellites color-coded by mission (recon, SIGINT, early warning)
→ GPS jamming zones with live severity overlays
→ Ukraine frontline updates every 30 minutes
→ 25,000+ ships via live WebSocket
→ 2,000+ CCTV feeds from NYC, London, Singapore

Right-click any point on Earth and get a full intelligence dossier.

The data has always been public.

Nobody bothered to aggregate it.

Until now.

100% Opensource.

Link in comments.

[MuleRun @mulerun_ai]

Introducing MuleRun 2.0.

Your personal AI, act before you ask.

It learns your habits, anticipates your needs, and works while you sleep — running 24/7 on your Personal Computer assigned to you alone.

No complex setup. Just talk to it.

[Priyanka Vergadia @pvergadia]

🤯BREAKING: Alibaba just proved that AI Coding isn't taking your job, it's just writing the legacy code that will keep you employed fixing it for the next decade. 🤣

Passing a coding test once is easy. Maintaining that code for 8 months without it exploding? Apparently, it’s nearly impossible for AI.

Alibaba tested 18 AI agents on 100 real codebases over 233-day cycles. They didn't just look for "quick fixes"—they looked for long-term survival.

The results were a bloodbath:

75% of models broke previously working code during maintenance.

Only Claude Opus 4.5/4.6 maintained a >50% zero-regression rate.

Every other model accumulated technical debt that compounded until the codebase collapsed.

We’ve been using "snapshot" benchmarks like HumanEval that only ask "Does it work right now?"

The new SWE-CI benchmark asks: "Does it still work after 8 months of evolution?"

Most AI agents are "Quick-Fix Artists." They write brittle code that passes tests today but becomes a maintenance nightmare tomorrow. They aren't building software; they're building a house of cards.

The narrative just got honest: Most models can write code. Almost none can maintain it.

[KK.aWSB @KKaWSB]

这哥们儿搞假直播，一天能赚几万块。

现在假照片和假视频已经是红海，假直播才是蓝海。

[chiefofautism @chiefofautism]

someone built a $96 3D-PRINTED MANPADS rocket that recalculates its mid-air trajectory using a $5 sensor and piano wire

its called Project Canard

it integrates with distributed camera nodes to triangulate airborne targets and update flight paths in real-time

it proves the barrier to advanced hardware has completely collapsed, moving precision weapons from defense labs to consumer garages

the entire launcher and interceptor frame is 3D printed in PLA and runs off a standard off-the-shelf ESP32 microcontroller

it even spins up a local Wi-Fi network so you can monitor live telemetry and arm the system directly from your laptop