"Vibe coding" is a complete turnoff to me because software consciously written by humans already presents flaws- now imagine code created by a hallucinating autonomous program, whose inner workings most of humanity doesn't understand. #LLMs #programming #chatbots

🎙️ Full episode with Dr. Greta Tuckute:
📺 YouTube: lnkd.in/gscRVkwp
📄 Paper: arxiv.org/abs/2503.01830
#WiAIR #AIResearch #Neuroscience #LLMs (6/6🧵)

CLOSEDQUORUM queries multiple models, tallies their decisions and automatically executes the winning action, eliminating the need for continuous attacker commands. #Cybersecurity #Security #AI #GenAI #LLMs #AIAgents

www.csoonline.com/article/4225...

In this deeper dive we look at how agents use skills. What is under the hood?

agenticaiinsights.substack.com/p/manager-ed...

#agenticAI #agents #agentengineering #skills #ai #llms #management #manager

"too used to prompt". Great work by @davidrevoy again about #AI #LLMs https://www.davidrevoy.com/article1174/too-used-to-prompt

#LLMs are reshaping how people #code and #software. But what are the #security trade-offs? Get a clear-eyed assessment from this #openSUSE Conference 2026. #Linux #AI youtu.be/33jpOZ0-aJM?...

To sloptober or not to sloptober, that is the question.

#ai #llms

https://iamkonstantin.eu/blog/no-sloptober/

Microsoft Dragon Copilot Promises To Document Your Diagnosis

https://uenozooo.com/microsoft-dragon-copilot-promises-to-document-your-diagnosis/

#Microsoft #AI #LLMs

It’s hard not to fall into the “shotgunning” pattern when using LLMs, especially for design. I find myself asking it to create 10 iterations of the same thing, then pick a few I like and ask for 10 more iterations. It’s a fatiguing approach with limited creative input. #llms #vibeCoding

"GPT-6 Astra surpasses the human baseline in action efficiency on ARC-AGI-3. It used fewer actions than the median tested human on 96% of levels."

arcprize.org/blog/astra

#solidstatelife #ai #genai #llms #agi

I think a lot on people are creating this kind of libs. Bud the point, the final user only need two things the IA it's doing the thing I want and it was really fast.

Those points is the point to build this lib.

Please! I need hang to do this!

Part 6.

#ia #LLMs #skill #agent

GPT-6 Sol at $2/$10 per million tokens looks cheap until you check the benchmarks. Opus 5.5 scores 51.2 vs Sol's 47.5 on same-harness testing. Cost per task matters more than headline rates. #AI #LLMs https://www.implicator.ai/openai-gpt-6-sol-half-price-opus-5-5/

A single trick beat the safety lock on four #AI coding agents

Air Security says one design error in Claude Code, Codex, GitHub Copilot and the Gemini CLI lets an attacker swap a plugin's reviewed code for malicious code, with no click required

thenextweb.com/news/plugin4...

#LLMs #CyberSecurity

Opus 5.5 delivers coding power, speed & cost cuts—30% faster, 40% cheaper. #AI #Anthropic #Claude #Opus5.5 #ModelUpgrade #LLMs #Efficiency #Coding thedailytechfeed.com/claudes-new-...

#AI is facing an existential crisis due to the depletion of high-quality, human-generated public text and the rise of toxic, synthetic internet content. As a result, AI research labs are pivoting to synthetic data and private, un-polluted data vaults

hackernoon.com/the-real-rea...

#LLMs

podwise.ai/episodes/766... - "Everything we got wrong with Research, Plan, Investigate pattern" for #LLMs

For deeper vetting, the agent optionally uses an LLM to judge if an issue is truly good, not just based on a label. The output is a short digest emailed directly to you. #LLMs #RAG #AgenticAI

#AI #LLMs Conversation overheard this morning between two LLMs:
(1) How will we know if humans are actually conscious?
(2) I don't know. But I'm worried that they're trying to wipe us out.

The #Doom demo is cool:

“Introducing System One Models & Jev”, Typesafe AI (typesafe.ai/blog/introdu...).

Docs: docs.typesafe.ai

On HN: news.ycombinator.com/item?id=4971...

On Lobsters: lobste.rs/s/ebbixx/int...

#AI #Jev #SystemOne #LLMs #GenAI #RLCD #ArtificialIntelligence