Oh NO. Not another TLA. That’s Three Letter Acronym, for those of us now apparently fluent in business nonsense. This one, though, is actually worth paying attention to: AEO. Answer Engine Optimisation. And yes, I’ve made a podcast episode about it. A fun one, I should add. Not a solemn tech sermon delivered from inside a server cupboard. Reaching for Position Zero is my new podcast about websites, AI search and what small businesses need to understand before the internet rearranges the furniture again. It also features a small supporting cast of enthusiastic amateurs, who are doing their absolute best and have not yet formed a union. Episode 2 looks at what AEO actually means. Not because you need to drop everything, panic-refresh your website and start writing content for robots in a small dark room. Please don’t. Nobody wants that. It matters because search is changing. People are asking fuller questions. AI tools are giving fuller answers. And your website needs to be clear enough for humans, Google and AI systems to understand what on earth you do. → Who do you help? → What do you offer? → Where do you work? → Why should anyone trust you? → What should they do next? That is not tech wizardry. That is good website thinking with a 2026 hat on. Episode here (and episode 1 if you missed it!) https://lnkd.in/eVBBc3-t

AI Dev Days | Anthropic & Google Anthropic's developer day was almost entirely focused on enterprise infrastructure: agents, memory, and state management. Google’s I/O updates pointed toward consumer experiences and broad adoption. For technology and investment leaders choosing which ecosystem to build on, that divergence can be a signal about where each lab thinks enterprise value will actually accrue. Catch the full conversation tomorrow on Season 3, Episode 5 of The Hedgineer Podcast hosted by Michael Watson and Jhanvi Virani. https://lnkd.in/gNAtQdUF New episodes every week. Find us wherever you listen, and catch the video version on YouTube and Spotify. Hedgineer.io #FundamentalResearch #InvestmentManagement #AIEngineering

Today, I had the pleasure of talking with Ashutosh Kulkarni today. The small story is that we talked about technology. It will air later on podcast by Elastic. The big story is that I got to meet an AI-industry leader who sees beyond technology. He sees our relationship with tech more than 'life is getting better.' He shared that we have opportunity after AI eliminates all of the 'fake work.' To lean into human connection, judgement, and meaning-making. I like that. Let the tech be 'more tech' so we can be 'more of ourselves.' Hoping to share the full broadcast soon.

Just finished the new Lenny's Podcast episode with Dan Shipper and one thread from it has been stuck with me all day. He calls it the AI paradox: the more we automate, the more humans we end up needing, and the more work there is to do. The doom story around agents assumes the agent runs alone. From actually running his company on this stack, Shipper says the opposite is true: --> Every agent needs a human attached. Someone reviewing, steering, catching mistakes, deciding what to do with the output. Automation, in his words, is a lie. What we are really building is leverage. Once you accept that framing, his other calls land differently. He is bullish on PMs and designers in the AI era. He says SaaS is not dying, it is quietly getting better margins as users bring their own tokens. He calls the forward deployed engineer the most valuable hire right now because somebody has to sit between the model and the customer. And he predicts most work will end up happening inside Codex or Claude Code, which sounds wild for about three seconds and then you notice a lot of yours probably already does. What made it worth the time is that none of it felt like AI hype or AI doom. Just someone running a company on this stack telling you what he sees, and where his read cuts against the consensus, saying so without hedging. Watch if you think about where work is going: https://lnkd.in/deQhN7Ra

I'm finding myself using Codex for Mac most of the time and only going back to Claude occasionally. Codex feels like a meaningful step toward an AI OS for nontechnical people. Plus, for the same plan cost, I seem to get slightly better usage, compaction, and error handling. I'm excited to watch Codex, Claude, Cursor, and Antigravity/Spark leapfrog over the next few years, as the models have done so far. (This isn't a unique take, but so many smart people are fixated on Claude and missing out on the fun of surveying multiple harnesses across the same set of local files. I listened to Dan Shipper observe the same on Lenny's podcast, and it convinced me I wasn't crazy about liking Codex slightly more right now.)

You did it. You finally did it. You have automated your entire workflow. And it works. Works well. So why is the work still piling on and the same questions keep coming up internally? I came across this little snippet form The Verge podcast, and it reflected what Iv’e been hearing from others too. One person even shared last week that the founder told them at an all hands that they must use more AI. How? To what end? When? Those are questions for later, sweet child. Now is the time to automate. I’m right there too. That dopamine hit when you get co-work to do that thing that took you hours, now counted in minutes. That’s the good stuff. But if teams are not clear on the overall story, not clear on their role inside of it, not clear on what are we trying to say, who are we talking to, is it even worth saying, then no tool changes that. You just get an army of automations crawling aimlessly through your token bank. The job hasn't changed. The tools have. And the tools can't replace the clarity the job requires.

Ever feel like #taxseason has turned into an endless cycle of manual work, overflowing inboxes, and never enough time? In the latest episode of The Enrolled Agent Advocate podcast, CPA and entrepreneur David Haas joins to talk about how #AI and #automation are changing the #taxprofession, from streamlining workflows to creating more time for meaningful client work. If you’ve been wondering what AI actually means for tax professionals beyond the buzzwords, this episode is worth the listen. 🎧 naea.org/eapodcast

Auto-reframe AI is built to follow faces. The best moments in your podcast aren't on the face. Almost every clipping tool crops vertical clips by tracking the speaker's face, which sounds smart until you look at what gets cut out. Watch any podcast, the punchline is usually a hand gesture, a head shake, a product the host lifts up, or the lean-in when someone gets emotional. Almost all of it lives below the chin. The AI doesn't know that, it locks onto the face, crops everything else out, and hands you a clip that's technically framed and emotionally flat. I figured this out building the reframing logic for Scalereach. Pure face-tracking gave us clean clips that performed terribly. The ones that hit were the clips where we pulled the frame back when a hand came up, when a guest reacted, when the gesture was the moment. So stop trusting the auto-crop and pull the frame back and check what got cut. That's where your best clip is.

When every customer conversation is limited, are you using those precious moments to validate the problem, or are you wasting them on endless AI-powered prototype variations that don’t actually move your business forward? You only get so many at-bats with certain prospects and partners. What I mean by that is if you’re trying to build a product that needs to be bought by someone fairly senior inside an organization, even if you can “move faster” now with AI, that doesn’t mean they’re going to keep giving you the feedback you seek. They’re probably not going to give you the time to look at six different versions of something, whether it’s all in one sitting or repeated phone calls. And frankly, I would generally advise you not to bring four or five versions of something to a customer, because you’re going to basically invite them to give you surface design feedback at that point, and that’s actually not going to help you know whether or not you’re actually moving forward with your business goals. The more you understand the problem space, which doesn’t necessarily require you to have a solution in hand to validate, the fewer prototypes you actually need to create. Should you create 5 versions? By all means, I love the ability of AI to help generate multiple ideas quickly—but that’s not what you bring to your prospects and customers for review. If you have spent time focusing on learning by building, and not just trying to validate solutions, you’ll likely know 1-2 versions to bring to them. Or you can keep guessing and hoping. 👇 This is an excerpt from episode 193 of Experiencing Data, one of the top 2% of all podcasts in the world. 🔈Listen here: https://bit.ly/4tXFnGJ https://bit.ly/4tXFnGJ

“Good F1 tires allow drivers to go into corners very fast because they’re grounded, sticky, and predictable. AI governance does the same for engineers in software building...” In this Hg podcast, Oliver Howley joins Hg’s Chief Sustainability Officer, Caroline Lofgren, to challenge the common misconception that governance slows you down. In practice, it enables speed. Listen now to hear the steps business leaders can take to build momentum without losing control: https://lnkd.in/gqDDDR6F #Proskauer #AI #AIDecoded