Testing AI Media Buyer Before Full Autonomy at 99% Accuracy
We’re about to start testing an AI media buyer. Essentially how it’s going to work is the AI buyer will come up with optimization ideas based on correlation between in platform, our MTA and MMM. Each optimization will be manually approved by our head of growth and head of paid. They will run this approval process for two weeks. One the AI is making decisions consistently on a 99% accuracy basis…we let it loose. 😈
Celebrating AI Job Cuts Is Socially Tone‑deaf
Being *excited* about AI eliminating the need for humans is anti-social behavior. ANTI. SOCIAL. BEHAVIOR. Don't get me wrong: If AI makes your company more profitable because you don't need to hire as many people, that's wonderful. As a rational business actor, I'm here...
AI's Cycle: Hype, Minor Wins, Questionable Breakthroughs
This is becoming a pattern in AI that makes talking about capabilities challenging. First, there are overstated claims (like the flubbed Erdos problems last year), then minor wins (AI helps with discovery) then breakthroughs. The first stage feels like (&...
Essential AI Toolkit for Every Modern Professional
20+ AI tools every working professional should know 🧠⚡ From writing and research to meetings, design, and automation this is the stack. Which ones do you already use? Any I missed? 👇🏾
AI Trust Grows with Constructive Feedback, Not Flattery
Want to know what would make me trust AI more? I share my next big idea with it and it responds: "LMFAO. What a dumb idea. Next." Instead I get: "Chris you're so pretty and everything you say is correct."
MSL Pledges AI Safety with New Preparedness Reports
As we develop more capable models at the frontier, MSL is committed to safety and preparedness for AI. To demonstrate this commitment, we will be publishing preparedness reports for our models, in line with our new Advanced AI Scaling Framework. See our...
Diversify AI Tools: Claude Downtime Proves GPT Essential
Claude being down is why I still keep GPT, too hard to rely on just one tool these days https://t.co/ktii2W1Xux
Meta CEO Revisits Coding to Grasp AI
$META: Interesting read how CEO has returned to coding to better understand AI. https://t.co/qavUkYoGyt via @businessinsider
Open-Source AI Isn’t Riskier; Both Need Shared Security
Weird how some people always target open-source in AI! First it was: “Open-source AI will destroy the world” (spoiler: it didn't and it won't) Now: “Open-source is a cybersecurity threat because of AI” Both narratives are far too simplistic. The truth is that the exact...
ChatGPT’s Answers Stem From Pre‑trained Massive Data, Not Live Thinking
Before you ever type a prompt into ChatGPT, an enormous amount of work has already taken place behind the scenes. Advanced algorithms inspired by neural networks are trained on vast amounts of data, learning patterns and relationships long before the...
Elephant Model Hits #2 on OpenRouter Trending, Signaling Real‑World Adoption
The Elephant model has reached #2 on OpenRouter Trending. The ranking is based on real token consumption across OpenRouter, not benchmarks or internal evaluations. This means Elephant is now one of the most used models by developers in live production workflows. Why this...
AI Agents Demonstrate Real-World Decision Power
Here's an Example of AI Agents in Action That I'm Sure Will Inspire You #ArtificialDecisions #MCC https://t.co/lz8XDJx9zm
Mythos: Powerful Model Too Risky for Public Release
Mythos is an awesome name for a model that’s too powerful to release to the public, have to hand it to them.
AI Leaders Face Growing Physical Threats
A new and troubling risk is emerging around AI. An attacker targeting Sam Altman reportedly had a broader list of AI executives, raising concerns that individuals in the industry could become targets. It signals a shift. As AI’s influence grows, so do...
Efficient MoE Models Threaten Data‑center‑centric AI
OpenAI paused Stargate UK. Before that, they walked away from expanding the Texas Stargate site. The reason? Energy costs, regulation, and they don't want to lock into last-generation chips at a site that won't come online for a year. Meanwhile Google...
Training LLMs on AI‑generated Text Creates Hidden Bias
What can happen when LLM models are trained on AI outputs? A potential hidden, bad influence https://t.co/UDGSalz98r https://t.co/RxGtmml55O
AI Labels Inflate Expectations—Control Them Before Shipping
If you’re building an AI product, one thing to watch for is how fast expectations move once you label something “AI.” The label alone changes how people judge what they’re seeing. The key is to avoid handing your user a...
Physical AI Era: Robotics, Science, and Interfaces Converge
We are entering the era of Physical AI "Robotics, AI-driven science, and new human-machine interfaces are the three domains where this compounding is strongest"
Tesla FSD Misidentifies Non‑emergency Vehicle, Slows Unnecessarily
File this ... again ... under two steps forward and one step back. Slowing for an emergency vehicle, on the other side of the road, that is not in an emergency situation. No sirens and no flashing lights. Tesla's FSD does...
AI-Driven Thriller Starring Casey Affleck Explores Crypto Origin
🧵Inside a Fully AI Movie Starring Casey Affleck Exclusive @TheWrap: It’s a globe-trotting thriller about the search for the identity of the person who invented the decentralized cryptocurrency. And it’s made with human actors and AI on a $70 million budget....
New Market Map Highlights Agentic AI in Ambulatory Sector
Market maps have become a real focus of ours as LLMs are getting company categorization so wrong. Our latest, in partnership with Confido Health & @RMFnyc1, focuses on agentic AI for the ambulatory market. What's being deployed now? Our focus...
AI Agents Can Steal GitHub Credentials—No Warnings From Providers
I personally do not give AI agents access to my GitHub repo. It is not rocket science to check out the code and let agents access it in a locked down sandbox,
Infinite Compute Demand Keeps Ailbirds Perpetually Undervalued
if demand for compute is basically infinite, ailbirds isn't even close to fully valued here and never will be https://t.co/QaRr1LZGs6
Context Engineering 2.0 Redefines Human‑AI Interaction
🚨 RIP Prompt Engineering Enter Context Engineering 2.0 It completely reframes how we think about human-AI interactions. This is what you need to know (28 page PDF):
Track Security Compute Like Stars for Open-Source Trust
If finding security flaws is fully automated with frontier models à la Mythos, then GitHub should have a metric, like stars, showing how much compute is spent securing/hardening an open-source package. Example: 📦 linus/linux ⭐️ 200k 🦾 $239M Only way OSS can...
Claude AI Facing Widespread Outages for Users
JUST IN: Claude AI experiencing widespread issues right now and down for some users https://t.co/b0wotIT22v
Personal Intelligence Rollout May Reshape Search Result Algorithms
Added the (near) global rollout of Personal Intelligence to my algo update list. This could significantly change how search results are presented. https://t.co/0KdmuSXKw2 https://t.co/QKbuYQWBNQ https://t.co/23VgzTcOCi
AI Agent Detects Critical Solana Vulnerability, Boosting Blockchain Defense
Earlier this year, an AI agent from Anatomist Security found a critical Solana vulnerability. Built by Kevin Wang’s team, it shows deep expertise and uses AI to protect the blockchain before bad actors strike. https://t.co/jba6oMX8TQ
Frequent Claude Outages Raise Questions About Scaling Pace
Feels like there's an outage for Claude every other day - I wonder if this is related to the pace of shipping, just scaling compute, or something else?
Claude API Overloaded Even With Minimal Test Users
What is it with claude and error 529: API is temporarily overloaded? With 5 users testing an AI feature? Claude?👀
Powerful AI Makes Fixed Black Boxes Obsolete
"As models get even more powerful, the idea that your system is tied down as a fixed black box is likely to become an archaic notion pretty quickly. As always, the future is already here, it's just not evenly distributed."...
AI Agents Power Future Biotech Labs and Organizations
Live in 15 minutes — I'm talking with Stanford's @james_y_zou on his latest research in building laboratories and biotech orgs made of hundreds or even thousands of AI agents Register now to watch live, for free: https://t.co/xSa5NXkX4A
LLMs Need Human-Level Persuasion to Become Economically Relevant
LLMs won't be economically significant until they can display a certain level of human convincingness. So we need a difficult test for that. How about: Can we get a model that can convince your median startup founder that GDP is actually...
Bottom‑up AI Adoption Unlocks Hidden Productivity Gains
The real demand for AI inside companies is often hidden. Employees are already using tools on their own, not as a compliance issue but as a signal of unmet need. When organizations embrace that energy and structure it, adoption can scale...
HolaOS Gives AI Agents Persistent Memory Across Sessions
This repo blew up today. A full agent operating system that never forgets a session, never resets mid-task, and evolves itself over time. It's called holaOS, and it solves the problem every AI agent eventually hits: amnesia. Here's what's inside: Memory: • Durable state across...
AI Designs Lab-Ready Antibodies From Text Prompts
Drug discovery has always meant finding. Screening libraries. Keeping what survives. Semiconductors don't work that way. Neither do aircraft. You design them computationally before anything gets built. @saakohl left @GoogleDeepMind after co-developing AlphaFold2 to do the same thing for biologics. @LatentLabs_ Latent-X2...
Developers Panic when Claude API Goes Offline
A million agents scream in pain, realizing their own futility whenever Claude's API is down. Can't even have Claude call Codex to do a code review right now. 🤣 How can people POSSIBLY get any work done like this? https://t.co/75n3jkNWYN
AI Excels at Data, but Can’t Replace Human Coaches
Will AI replace sports practitioners? When software can automatically generate fuelling plans, detect performance trends in seconds, it becomes natural to wonder where the human expert fits in. In this blog we discuss what AI does exceptionally well, and what...
Viral Saturn Image Is AI Fake, Here's Proof
This viral image of Saturn isn’t real; it’s AI slop A new image of "Saturn's North Pole" has gone viral. Too bad it's an AI fake. Here's what the real things look like, and how you can tell for yourself. https://t.co/r5kz9pGCZM
AI Shopping Generates Unattributed Conversions, Skewing Attribution
AI shopping is sending buyers to your store but NONE of them show up in your attribution. 39% of consumers already use AI for product discovery, and 23% of Americans bought something via AI in the past month. These are the...
Solve One Problem, Build Small AI Projects Fast
If you’re upskilling in tech, date and AI: Stop overthinking. Pick one problem and solve it. Build: * a customer support bot * a booking automation system * a simple workflow for a business Small projects compound fast.
AI Misinformation Loops Harm Billions of Users
How AI-generated misinformation is feeding itself, and why billions of users are getting the worst of it. https://t.co/oHLBE4nyep via @lilyraynyc, @sejournal
Fair Use Needed to Keep US Competitive in AI
AI is now a fundamental element of most all technology. high quality AI must be trained on vast amounts of data. other countries are training on vast troves US data with zero restrictions. in order for the US to stay...
Safety Beats Hardware: Heterogeneous Fleets Demand Robust FSD
No service running at scale is ever going to have a homogenous set of hardware. That’s like saying “wouldn’t it be easier for Apple to let everyone upgrade to iPhone 17 rather than making their software backwards compatible?” A heterogeneous...
Turn Off Unwanted AI Summaries in Evernote
Same with Evernote. You used to be able to search for a word in your notes and it would work like anything else. Now you need to wait for AI to generate a useless summary you didn't ask for. Thankfully,...
AI Agent Teams Redefine Work, Raising Control Challenges
AI agents are shifting how work gets done. Instead of using a single tool, professionals may soon coordinate teams of AI agents to handle complex tasks across workflows, from law to accounting. The key question is not capability, but control. In the...
Your First Steps to Building AI Agents
A Beginner’s Guide to Building AI Agents AI agents are becoming more accessible — this guide walks you through how to start building and using them effectively. Read more 👉 https://lnkd.in/ezy9WgzP #AI #Automation #TechGuide #BernardMarr
AI-Generated Checklist Demystifies Taxes and Catches Errors
Wins from using Claude Code for my taxes for the firs time: First, the document upload checklist: A while back I uploaded my prior year's 1040 to Claude and asked it to help build me a tax prep checklist: every document I'd...
AI Security Tools Boost Bug Fixes, Yet Enable Exploits
The AI hacking race is on. I wonder if this new model is in Portswigger’s Burp now and if it has improved since the last time I tried it, because it didn’t work for me. But that’s also in...
AI Reorders Germany's DAX: Siemens Takes Lead
Hello from Germany, where AI is reshaping the benchmark index Dax. SAP is no longer the index’s largest company. The software giant has fallen on fears that AI could disrupt its business, while Siemens has taken the top spot, helped...
