Eric Schneider on AI's Impact on Product Development

New episode of the AM Podcast is live! In EP3, I sat down with Woosuk Kwon — co-founder & CTO of Inferact and creator of vLLM — to explore LLM inference from a user-centric perspective. There's no shortage of content covering the implementation side of LLM inference. But two questions I keep coming back to are: how are emerging applications reshaping AI infrastructure? And how do you make the right assumptions when designing complex systems? In this conversation, we cover: - Woosuk's first PhD project and how his prior experience shaped his taste for infrastructure work - How vLLM started and what made it take off - How to prioritize feature requests for a widely-used open-source project - How emerging applications are reshaping AI infrastructure - What's on the frontier: streaming requests, continual learning with RL, on-device inference, and more - Why to start an AI inference startup - Whether today's coding agents could build something like vLLM - And more! This conversation answered a lot of questions I personally have. Whether you work on user-facing applications or deep in the infrastructure stack, hopefully it can offer you something new! 📺 🔗 https://lnkd.in/gfKByU9Y

80% of AI projects die in the "Last Mile." Let’s fix that. Anyone can build a flashy AI demo over the weekend. But converting that sandbox prototype into a secure, scalable, and affordable product for the enterprise? That’s where the real work begins. In Episode 3 of the Just Do-it Right podcast, we are thrilled to welcome Brian Gagne, CTO and Senior AI Engineer for Kief Studio. We’re tearing down the "Last Mile" graveyard and discussing: 🪦 Why 80% of AI initiatives fail to reach production. ⚙️ How MLOps is the missing link between a cool demo and a reliable product. 🚀 The operational playbook for scaling AI safely across both SMBs and Enterprises. Stop building prototypes and start deploying products. 🎧 Listen to the Episode Three Preview with Brian on Spotify: https://lnkd.in/grmZ9tVu Official launch of Just Do-it Right Podcast Season One is April 3rd, 2026..... #AI #MLOps #TechLeadership #EnterpriseAI #JustDoItRight #KiefStudio

Excellent discussion with Johnathan. The reality of AI in work, life, and how it plays out in practice. Thanks again for the journey

Most businesses don’t have a people problem. They have a systems problem. AI is just exposing it faster. A quick insight from the Ctrl+Alt+Dev podcast on how software evolution is changing the way businesses scale. Link in comments 👇

AARON Awards – Podcast Series This week we chat with Arlyn Panopio from Envato, one of the partners of the AARON Awards and someone with a front-row seat to how AI is reshaping the global creator economy. A few themes from the conversation stood out: ⭐ The barrier to entry has dropped dramatically, but that makes creative judgement more valuable than ever ⭐ AI isn’t replacing creators, it’s expanding who gets to create — and how fast they can move ⭐ The shift isn’t just in tools, it’s in expectations around speed, scale and output ⭐ And importantly: platforms like Envato are seeing firsthand how this is changing what clients and brands demand from creative work What makes this conversation interesting is that it comes from a platform perspective. Not just how work is made, but how it’s distributed, discovered and valued at scale. Exactly the kind of thinking that’s shaping the work coming into the AARON Awards. Because the goal isn’t just to celebrate AI being used. It’s to celebrate how it’s used to push creativity forward. 🎧 Worth a listen 👉 https://lnkd.in/g-GfNfuk #AARONAwards #CreativeAI #AIinAdvertising #Envato

Most agencies are rushing to cut costs with AI. This Is Flow is doing the opposite 🧠 In the latest episode of Mind the Model, Catherine Rushton, CSO at This Is Flow, explains why a people-first approach outperforms the algorithm. Three things worth stealing: 👥 Let junior staff lead their own automation audits. They know where the grunt work lives better than leadership does. 🎙️ Transform dense documents into short podcasts with NotebookLM. Crush used this approach for the Media Federation of Australia Awards judging, resulting in real engagement from the judges. 🎤 Voice memo your strategy on your commute. Crush uses Gemini to turn a stream of consciousness into sharp client briefs before she's even at her desk. Crush's take: speed up the draining tasks so you can slow down on the ones that matter ⚡ 🎧 Full episode in the comments. Link in the comments. #MarketingInnovation #GenAI #ModernMarketer #AudienceIntelligence #AgencyLife #MindTheModel #MindtheModelPodcast #PeopleFirst Nathan Guerra

Most people think spotting an AI unicorn is about timing the market - they're wrong. It’s about seeing what others can’t yet see—and having the conviction to move before it’s obvious. For our 100th episode (one-hour special) of the LeaderbookAI Podcast, I sat down with Marina Davidova, Managing Partner at DVC, who has backed 4 AI unicorns in just 4 years. What struck me wasn’t just the outcomes. It was how she thinks. While everyone else is chasing signals—traction, metrics, hype—Marina is looking at something deeper: Has this founder already been through failure and come back sharper? Do they think like an engineer—able to deconstruct and rebuild systems? Do they have the energy to keep going when the market turns against them? Because at the earliest stage, there’s almost nothing else. No perfect data. No certainty. No guarantees. Just people. And in a world now flooded with AI tools, faster diligence, and more capital than ever before… The edge isn’t more information. It’s better judgment. We also got into something most people aren’t talking about: How venture firms themselves are using AI— not just to invest in companies, but to screen deals, surface signals, and make decisions faster than ever. And yet, even with all of that… The final decision still comes down to one thing: 👉 Who do you believe can build through uncertainty? This conversation is for anyone: building in AI investing in AI or trying to understand where real value will be created next 🎧 Episode 100: How do you spot AI unicorns before everyone else does? Link in the comments below. If you’ve been following the journey—thank you. This one is special.

Every day brings a flood of new articles, podcasts, tools, and “best practices,” all promising to help us move faster and stay ahead. But the reality is, keeping up already feels impossible… and it’s only getting harder. This creates a new kind of pressure inside our organisations. Not just to deliver more, but to learn faster, adapt faster and continuously re-skill in real time. So as leaders, the real question is: How do we help our teams operate effectively in an environment where the rate of change is outpacing their ability to absorb it? - https://lnkd.in/eg2ysPj5

Dismissing AI is no longer an option. Rami Alhamad, Founder of Alma and PUSH (acquired by WHOOP) breaks down why in this episode of the AGI Ventures Canada podcast. Technology is moving so fast that what was true six months ago is already outdated. Developers who treat AI like a junior dev, building systems to manage and evaluate it rather than ignoring it, are the ones who will scale. Watch the full episode on YouTube. https://lnkd.in/esFWCwCA

Self-healing codebases are here. This week is the first week where every bug report and every error on our prod instance at Alma triggers a headless claude code instance that audits, proposes a fix, and makes a commit. I wake up to 4-5 commits that I review, and then push to production.