CEO AI Adoption Insights with Samar at Temporal

This is a fantastic article from Jani on something every developer and engineering leader needs to understand right now. In one weekend, OpenAI doubled its flagship model price while DeepSeek released open-weight models at a fraction of the cost. The middle tier - where most coding agents and consumers have been living - is hollowing out.

A few people asked a fair question after my last post. Okay, but how are we growing media properties in 2026? What actually changed? I'm happy to share. We have three sites, so there are three different answers, but one shared habit. We're obsessive about our target audience. Roadmap is the rocket ship, and it runs on a simple mandate. Keep every Roadmap current, and launch new ones the moment demand emerges. We moved fast on Claude Code, OpenClaw, and a forward deployed engineer roadmap. Then we added premium tools that serve the same core users, like the AI tutor, a teams plan and Lesson Packs. The New Stack rebuilt what it covers. We cut way back on evergreen explainers and on second-day analysis that landed late. In their place: breaking news written for our core audience, second-day stories on the first day, and timely technical analysis and tool comparisons. We write for developers and engineers building with AI, and we do it with a sense of urgency and authority. The traffic validates this approach. Breaking news sometimes leads, but for the most part, ultra-timely, in-depth analysis carries about half of our top traffic, and people stay on it roughly twice as long as breaking news stories. Towards Data Science grew on steadier ground. Ludovic Bénistant and Ben Huberman run it with a consistent hand, and we pay writers through a fair, competitive formula tied to engaged readership. In short, good work, paid fairly, brings more good work. That's most of the story. None of this is a secret formula. Know exactly who you're building and writing for. Give them something timely and useful. Pay the people who make it. Then do it week after week. I'm lucky to work with teams that do this every day. And we're hiring to do more of it. If this is your kind of work, DM me.

Most executives have valuable insights. Very few have time to write about them. After decades covering technology and business as journalists, we’ve seen this repeatedly. Great founders and CEOs often have strong perspectives on their industries — but turning those ideas into clear articles, commentary, or LinkedIn thought-leadership posts usually falls to the bottom of the to-do list. So we decided to help solve that. Mike Brennan (MITechNews) and Matt Roush (formerly of the Great Lakes IT Report) have launched an executive ghostwriting service designed for busy leaders who want to share their expertise but don’t have time to write. Between us, we bring more than 70 years of journalism experience covering startups, technology, innovation, and business. We’re also co-hosts of MITech TV, a streaming program covering Michigan’s innovation economy, so we understand how strong executive ideas translate across articles, LinkedIn posts, interviews, and video. Matt also brings extensive radio and voice-over experience, which can help executives expand into podcasts, commentary, and broadcast content. Our process is simple: • Short interview with the executive • We turn the conversation into a polished article or LinkedIn post • Written in the executive’s voice • Ready for publishing or sharing Once completed, the article can also be shared across our LinkedIn networks, which reach thousands of technology executives, entrepreneurs, investors, and media professionals. Important note: This is not pay-to-play journalism on MITechNews. The content is created for the executive’s own platforms and social channels. If you’re interested in building stronger thought leadership around your expertise, feel free to reach out. More details: https://lnkd.in/g93VPNXr — Mike Brennan Editor & Publisher, MITechNews Co-Host, MITech TV — Matt Roush Former tech editor, WWJ Newsradio 950 Great Lakes IT Report Co-Host, MITech TV

Over at IT Brew, Eoin Higgins sat down with the one and only Ed Zitron to talk about the potential for an AI bubble, the ROI of AI, government subsidies of the technology, and much more. It's a huge chat -- so huge that we broke it into two parts. If you're not one of the (many) readers of Ed's newsletter, or you haven't seen his multiple appearances on TV, etc., he's hugely skeptical of AI's societal potential. "We are very much in the desperation phase of the #AI bubble, because even with incremental improvements to models, no one can measure the actual ROI of AI," he told Eoin. "The actual capabilities—forget benchmarks, which are rigged for the companies to literally train the models to perform on—the actual things that these models can do are basically the same as they were a year ago. They might be better at coding, but they are still mindless, thoughtless machines churning out code with no intention, and they have categorically failed to expand new industries. They just haven’t." Read more at ze link! https://lnkd.in/gxw_mi2e

My adventures in AI: As I wrote last week, I've been playing with Writer after the company set me up with the software, which is designed for enterprise users, not a solo operator like myself. Last week, I talked about its editing skills. This week I want to look at its ability to build agents, what's akin to Claude Cowork. To do a fair test, I used a similar prompt to the one I used with Claude Cowork to send me a custom tech newsletter every morning. It did a great job of taking my prompt and turning it into a Playbook where I can see exactly what it's going to do. * On the plus side, it sends me a newsletter every morning to my work and personal emails. So far that has worked flawlessly. Claude has trouble emailing it and puts it in my drafts folder, instead. * It works regardless of whether my computer is open and awake. Claude won't run unless my computer is open. * It's really nicely formatted and easy to read. It does a summary of what the big trend was over night and gives me a nice variety of news. * But it requires constant tweaking. The first day, it delivered a headline that IBM had bought Confluent. That happened in December. I pointed that out, and it adjusted the Playbook. * I have to keep telling it to use good sources or it will choose some pretty random websites. * It still has trouble with the idea of fresh news and I have to instruct it every day to make sure it only gets new stories. It has trouble understanding that I want it to be what happened over night. In general, it works well, but it's further proof to me that it's hard to get an agent to do exactly what you want without constantly giving it feedback. This is a pretty simple exercise to check the news and deliver the top stories to me every morning. But a week into it, it's still not quite right. Now imagine what it would be like getting an agent to understand and work through a much more complex enterprise workflow. This isn't an indictment of Writer, it's a reality check on agents in general. They take more work than people expect, and they don't just work out of the box.

The coldest winter I ever spent was a summer in San Francisco. CTOStudio is back in Silicon Valley this week, both for the sweater weather and for our next round of interviews. I'll be conducting 15 on-camera sessions with CTOs, CPTOs, and engineering leaders across a myriad of industries. In aggregate they're leading thousands of devs and shipping billions of dollars worth of product. So understanding how AI tools are impacting their orgs, for good and for ill, helps me understand the downstream effects. The pace of AI news is staggering, and the mood outside the valley can best be described as anxious. I look forward to sharing what the mood inside the valley is like as a counterpoint. Before each weekly session I've started reading a different AI-centric tome. This week's is The Coming Wave by Mustafa Suleyman. His core warning is what he calls the containment problem: once a technology this general and this fast is loose, steering it gets very hard, and the unintended consequences are exactly the ones nobody priced in. That's why those downstream effects are worth chasing in person. The leaders I'm sitting down with this week are the ones actually trying to steer it, inside real teams, with real stakes. Forewarned is seldom forearmed... but I'll do my best.

If you are an early stage startup founder trying to find your story, this podcast from my former TechCrunch colleague Walter Thompson is absolutely a must-watch for you. Walter has built a great property with Fund, Build, Scale (great name) and this episode is jam-packed with practical advice on how to build and tell a concrete story that travels from investors to customers to potential employees, and yes, to reporters like me. This resonated with me so much because before I launched FastForward, I had to communicate my vision to PR people, who might have known me from my TechCrunch days (and before that), but who had no anchor for FastForward because it didn't exist yet. I had to tell my story and get them to put their executives in front me on an act of faith. Granted it's one that came with an existing reputation, but there was no publication yet, nothing to show the team, and that was a risk. I had to tell them what my vision was. That's what all startup founders have to do, and that's what Walter articulates so well in this episode. https://lnkd.in/e28u7Ypa

Today's adventure in AI: I wanted to look at the latest research on AI in the enterprise to confirm my current understanding or to see if it has evolved and things are finally moving. I fed a fairly complex research prompt about the latest 2026 AI reports into three models: Perplexity, Gemini and Claude. Perplexity and Claude both asked pointed follow‑up questions, including whether I wanted to include vendor‑sponsored reports (which can introduce bias) and who the audience for the report was. Gemini provided less depth and didn’t ask many follow‑up questions, beyond offering to slice results by industry such as finance or healthcare. Perplexity did a solid job on the research and surfaced relevant 2026 studies, but struggled to generate a downloadable PDF. Claude asked the most detailed follow‑up questions and generated three reports with increasing levels of detail. It had no problem producing downloadable PDFs. Bottom line from all three models, based on the 2026 reports they pulled: the research still matches my writing—the hype is way ahead of the reality. I plan to dig into the reports themselves (I’m not taking the LLMs’ word for it), but they provide a good starting point, with links to primary sources for deeper research and analysis. Here's a piece I wrote last week about agent hype in case you missed it. https://lnkd.in/eBHDCGM9

A serious investigation from my Technical.ly colleagues Alice Crow and Katie Malone deserves attention today. Content warning: The story discusses child sexual abuse material and sex-offender registry laws. It does not include graphic descriptions. Following an anonymous tip, Technical.ly identified a Pittsburgh tech founder and CTO who had previously been convicted in a federal case involving child sexual abuse material. https://lnkd.in/ej_wzbaJ The reporting found that he appeared to be using a LinkedIn alias (a slightly different spelling of his last name) that was not listed on his public sex-offender registry profile. That matters because Pennsylvania’s Megan’s Law requires registered sex offenders to keep certain information current, including aliases. That's how our society has decided to navigate these circumstances. After Technical.ly contacted state police and later reached out to him for comment, the alias was added to his registry page the same day. Pittsburgh police are now handling the compliance investigation. This is hard, sensitive reporting. It is also a reminder that technology and economics journalism cannot just be boosterism. Technology is a tool society uses. So it reflects our ambition, creativity, wealth and power. Our failures too. Years ago, I worked on a different reporting project about a company that helped convicted sex offenders bury search results. A source challenged me then: Why are you chasing broken men trying to rebuild their lives? I believe in rehabilitation. But two things have to come first. The first priority is survivors. Always. Second, society has created rules for disclosure and accountability. Those rules do not necessarily mean someone can never work again. But they do mean compliance and transparency are not optional, especially when someone is taking a public-facing leadership role in a tech startup. That’s why this reporting matters. Read Alice and Katie’s investigation here: https://lnkd.in/ej_wzbaJ