Clever take on AI benchmarks by Angela Aristidoua professor at University College London and a faculty fellow at the Stanford Digital Economy Lab. In an article on MIT Technology Review @techreview she makes the case that current AI benchmarks are broken because they test models in isolation rather than in real-world workflows. Instead of raw scores, we need metrics that evaluate "human correction load," error detectability, and overall team productivity to reflect how AI actually performs as a collaborative tool. Read another way, hiring the "smartest" person who is difficult to manage is like using a powerful but "black box" AI; the time spent correcting their social or procedural errors (the "downstream damage") often outweighs their individual output. Ultimately, both AI evaluation and hiring go wrong when they prioritize isolated capability over collaborative reliability. technologyreview.com/2026/03…

Anyone else having this fever dream?

Great conversation with Asha J. Palmer, AI governance, ethics, and compliance leader at Skillsoft about what responsible AI actually looks like in practice. Asha brings both the legal compliance lens and the human one, and the result is a great discussion about what it means to stay relevant in an AI-powered world.

Here are the top euphemisms AI companies will use to soften the reality of automation and job displacement: Efficiency & Process Language * Streamline * Optimize * Augment workflows * Accelerate * Transform processes * Human-Centered Framing * Augment (workers, not replace them) * Empower employees * Free up" humans for higher-value work * Upskill / Reskill * Co-pilot (implying humans are still flying the plane) Abstract / Corporate-Speak * Scale (operations without headcount) * Drive efficiency * Unlock capacity * Reallocate resources (resources = people) * Right-size the workforce Innovation Framing * Future-proof * Modernize * Digitally transform * Enable next-generation workflows The Subtle Ones * Reduce friction (the friction is often a human step) * Automate the mundane (until the mundane is your job) * Handle repetitive tasks (entire roles are repetitive) * 24/7 availability (no lunch breaks, no salary) The marketing team will frame AI as a collaborator or enhancer rather than a replacement, and wrap displacement in language that sounds like a promotion. "Your job isn't being eliminated — it's being evolved."

The Human Conversation continues to roll along with new weekly releases. This latest episode features the always compelling @EvanKirstel - B2B thought leader and top technology influencer who helps clients grow their social media audience. As you can see in this excerpt, I got in to the existential crisis we faced the first time we saw ChatGPT a little over three years ago. And how our team at Solutions Review and InsightJam.com has been able to pivot our content and authority into online events, video production and community building. All the while connecting with a host of new experts and organizations that share our concern about the human impact of AI.

Read this quote from Matt Shumer the co-founder and CEO of OthersideAI, an applied AI company building some of the most advanced autocomplete tools in the world. "I am no longer needed for the actual technical work of my job. I describe what I want built, in plain English, and it just… appears. Not a rough draft I need to fix. The finished thing. I tell the AI what I want, walk away from my computer for four hours, and come back to find the work done. Done well, done better than I would have done it myself, with no corrections needed. A couple of months ago, I was going back and forth with the AI, guiding it, making edits. Now I just describe the outcome and leave."

Online news will experience the double-whammy with AI impacting its workforce. 1. AI can easily replace the content creation. And, 2. With AI now providing answers at the top of search returns, the "zero-click" result will crater organic traffic.

There is a wildfire just over the hill. Right now you can only see the smoke. "Lawmakers don't get it or don't believe it. CEOs are afraid to talk about it. Many workers won't realize the risks posed by the possible job apocalypse — until after it hits."

AI will easily remove the bottom rungs of many traditional white collar career ladders; the junior coder, the sales development rep, the associate marketing manger and many more are all jobs where companies will employ AI agents in the next couple of years. Guidance counselors and universities will be slow to react.

It is advancements like this that will disrupt all of traditional education. The advantage of vast institutional knowledge no longer exists for academia. Creating a way to deliver tutelage in a more dimensional way will begin to eliminate the in-person human delivery advantage. And it may well increase focus for the student and speed learning; so time for education is metered out and measured at the individual level and not throttled by the average of the cohort. In short - smart, motivated kids won't need to sit - bored to death - while the learning bureaucracy delivers its mediocre curriculum to the others.

I am genuinely excited for that kid that wants to tell a story in their head that only they can see. Movie making will go back to a simpler time, when the story mattered more than the return on investment.

It doesn't happen all at once, but it is already happening. And we will be slow to react.

Here are the verbatim elevator pitches of a few AI Agent startups direct from @ycombinator (gotta love #15 - hoo boy) Agents that automate enterprise software processes Agents that find your perfect customers Agents for Autonomous Influencer Marketing Agents to help find in-market companies via deep researched signals Agents for dental operations Agents for Accelerated Clinical Trial Design Agents for healthcare, powered by clinical and wearable data. Agents that earn consumers dollars when they shop, travel, and live. Agents for customer engagement on WhatsApp, Voice, Chat & SMS Agents for sales and support teams, starting with training and QA Agents for QA testing Agents for Sales Operations Agents to Replace Restaurant Managers Agents for recruiting firms Agents to Kill Salesforce Agents Automating the $400B Tech Consulting Market Agents for Revenue Intelligence Agents for regulated industries Agents for Audit Agents to automate corporate transactions Agents for Marketing Compliance Agents for life sciences research Agents for fraud detection and investigations Agents that troubleshoot production alerts for software engineers Agents for B2B SaaS Finance Teams Agents for AP and AR in Logistics Agents for private markets that accelerates deep financial work Agents for precise document analysis Agents for healthcare administration

YC Companies building Agents 41 (37%) of out the 111 Spring 2025 Y-Combinator companies are building agents