📬 This is the companion episode guide to 5x Faster Code. 18 Months to 4.
Subscribe to get the full briefing →
Episode Guide: 5x Faster Code. 18 Months to 4.
Companion to the Sunday, March 15, 2026 edition of Transformation Brief: AI & Technology
This edition covers 12 episodes spanning AI in software development, AI ethics and governance, AI data sourcing, AI in design, AI ROI measurement. Below you'll find detailed breakdowns of every episode referenced in today's briefing — including key guests, standout quotes, and links to listen.
AI Impact Beyond the Hype: From Agentic Coding to Ethical Battlegrounds
This week, we’re cutting through the noise to bring you the signal on how AI is truly recalibrating our operational realities. Dive into the surprising shifts noted by Notion’s co-founder as he ditches coding for agent management, Canva’s strategy for user acquisition via free AI tools, and the looming ethical battles influencing AI deployment in everything from data analysis to national security. Get ready to rethink your AI strategy—because the next wave isn't just about adoption, it's about adaptation and principled deployment.
No Priors: Artificial Intelligence | Technology | Startups — "From Coder to Manager: Navigating the Shift to Agentic Engineering with Notion Co-Founder Simon Last"
Runtime: 29 min | Host: Sarah Guo | Guest: Simon Last
For the engineering leader watching their dev team’s roadmap stretch endlessly: This one's for you. Notion's co-founder reveals how he pivoted from coding to "agent management" and the cascading impact on product development and engineering velocity.
Simon Last, co-founder of Notion, unpacks the company's aggressive, iterative pivot into AI, moving from basic writing assistance to a long-term vision of autonomous agents manipulating Notion tools. The core insight? The engineering team rewrites its AI harness every six months, and Last himself has stopped writing code, effectively becoming an "agent manager." This isn't just a tech upgrade; it's a fundamental shift in how software gets built, scaling individual output through agent-friendly APIs and custom, empirical retrieval strategies for data integration.
"Our goal was to create the best tool for humans to directly perform their work. And then now the goal is to create the best tool for humans to manage agents, to do the work for them." — Simon Last
Connects to: AI in software development, Shift from humans doing work to humans managing agents, Coding agents increasing ambition
The Neuron: AI Explained — "24 Billion AI Uses Later: What Canva Learned About the Future of Design"
Runtime: 55 min | Host: Corey Knowles | Guest: Danny Wu (Canva), Grant Harvey (The Neuron), Danny V M (Canva)
For the product or marketing exec battling design homogenization and seeking new customer acquisition channels: Canva's Head of AI Products spills the beans on their 24 billion AI usages and what it means for design, personalization, and user growth.
Canva's Danny Wu shares insights from 24 billion AI uses, revealing how the platform has redefined design beyond static images to editable, layered outputs. Their "Creative Operating System" approach integrates AI deeply, including offering free AI tools on external platforms like ChatGPT and Claude—not as cannibalization, but as an “new SEO” strategy for user acquisition. The focus shifts from preventing design homogeneity to enabling deep personalization and style training, highlighting a future where AI empowers greater creative diversity, not less.
"You shouldn't be afraid to cannibalize your own product." — Danny V M
Connects to: AI in design, Editorial layers in AI image generation, AI assistants becoming 'the new SEO' for user acquisition
The AI in Business Podcast — "Building a Virtuous Cycle of Analytics in Global Enterprises - with Barry McCardel of Hex"
Runtime: 39 min | Host: Daniel Faggella | Guest: Barry McCardel
For the CFO or Head of Data trying to prove ROI and escape "dashboard hell": Discover why internal sentiment matters more than logs, and how leading with experimentation over grand prophecies can actually get AI projects off the ground.
Barry McCardel, Co-founder and CEO of Hex, debunks the myth of solely dashboard-driven data strategies, arguing that they often create more questions than answers and lead to "tool creep." He posits that the true ROI of a data team is found in internal sentiment and collaboration, not quantitative logs. McCardel advocates for a "crawl, walk, run" approach to AI, emphasizing fundamental data organization and rapid, bottom-up experimentation over top-down, big-bang deployments often plagued by 95% failure rates.
"The problem is always like, dashboards raise more questions than answers." — Barry McCardel
Connects to: AI ROI measurement, High experimentation mentality in AI adoption, Data ROI in internal sentiment vs. quantitative logs
The AI Daily Brief: Artificial Intelligence News and Analysis — "The Debate Over Anthropic’s New Product: Price or Existential Dread?"
Runtime: 26 min | Host: Nathaniel Whittemore | Guest: Nathaniel Whittemore
For the executive grappling with the disruptive potential of specialized AI services and the market's response to pricing labor-replacing tech: This debate peels back the layers on Anthropic’s $15 per pull request code review and what it truly signifies for pricing power and industry disruption.
Nathaniel Whittemore dissects the uproar over Anthropic's Claude Code Review pricing, arguing it’s a canary in the coal mine for broader industry disruption. The $15-$25 per pull request cost isn't just about price; it’s about the market’s struggle with valuing AI that replicates specialized, high-cost human services. This controversy highlights the existential question for developers and other knowledge workers: as AI foundational models internalize features and services, what does it mean for pricing power, industry consolidation, and the future of human-led workflows?
"This marks the beginning of the end of the subsidized inference era. It will only go higher. I think we are just beginning to grapple with what the full bore cost of AI when fully utilized is going to look like and what it means for the structure of organizations." — Anonymous 4o account on Twitter
Connects to: AI's impact on white-collar jobs, Consolidation around enterprise AI stack, Existential dread in AI development
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) — "Agent Swarms and Knowledge Graphs for Autonomous Software Development with Siddhant Pardeshi - #763"
Runtime: 76 min | Host: Sam Charrington | Guest: Siddhant Pardeshi
For the CTO or Head of Engineering aiming for hyper-scaled autonomous development: Discover Blitzy's unconventional approach to agent swarms and knowledge graphs that promises 5x faster software development by treating "code as a commodity."
Siddhant Pardeshi, Co-founder and CTO of Blitzy, challenges the conventional wisdom of autonomous software development. He argues that "code is a commodity," with the real challenge lying in "code acceptance" (security, standards, maintainability). Blitzy tackles this by deploying hyperscaled agent swarms orchestrated by a custom hybrid graph-plus-vector approach for efficient codebase navigation. Unlike traditional multi-agent systems, Blitzy's agents dynamically assign personas and tool selection, yielding 5x faster development—transforming 18 months of work into 3-4 months.
"code is a commodity... Getting AI to write code is very easy. Getting any code is easy. Getting code that follows your standards, codes, that code that is really good, it goes, that is secure, code that is ready for production is a completely different story." — Siddhant Pardeshi
Connects to: AI in software development, Autonomous software development vs. AI-assisted coding, Dynamic agent persona design
Me, Myself, and AI — "An Industry Benchmark for Data Fairness: Sony’s Alice Xiang"
Runtime: 34 min | Host: Sam Ransbotham | Guest: Alice Xiang, Shayan
For the compliance officer or risk manager navigating the murky waters of AI ethics and data bias: Sony AI's Alice Xiang reveals a practical framework for ethical data sourcing for AI, cutting through the "data nihilism" to establish new industry benchmarks.
Alice Xiang of Sony AI addresses the pervasive challenge of data bias in computer vision with PHOEBE, their new ethically-sourced benchmark. She combats "data nihilism" by demonstrating that responsible data collection—with consent and compensation—is achievable, albeit more difficult. The discussion highlights that while AI adoption is accelerating into corporate KPIs, the practical implementation of AI ethics and human oversight lags significantly. Moreover, the conversation draws surprising parallels between classical operating system design principles (like isolation) and critical elements for designing agentic AI systems.
"AI isn't just a technology shift, it's a leadership test. The real challenge isn't adopting AI, but knowing how to apply responsibly, productively and at scale." — Sam Ransbotham
Connects to: AI ethics and governance, Data fairness in computer vision, Ethical sourcing of data for AI models
"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis — "Bioinfohazards: Jassi Pannu on Controlling Dangerous Data from which AI Models Learn"
Runtime: 103 min | Host: Erik Torenberg, Nathan Labenz | Guest: Jassi Pannu
For the biotech investor or a board member concerned about biosecurity risks from AI: A Johns Hopkins professor lays out the urgent need for data controls on functional biological data, and why "bioinfohazards" are not just theoretical, but a pressing reality.
Jassi Pannu, Assistant Professor at Johns Hopkins, illuminates the escalating biosecurity risks posed by AI's advancements in biological research. She differentiates between the abundance of easily accessible genetic sequence data and the critical need for data controls on "functional data" for viruses with pandemic potential (e.g., transmissibility, immune evasion). Pannu advocates for a "Biosecurity Data Level" framework, akin to biosafety levels for wet labs, to restrict sensitive data while maintaining open science, especially as AI agents become adept at discovering and exploiting signal-rich data for viral design.
"The most abundant biological data that currently exists is sequence data. We are currently swimming in petabytes and petabytes of sequence data, a lot of which we don't know what it does or the functions of those different sequences." — Jassi Pannu
Connects to: AI data sourcing, Gain-of-function research risks, Biosecurity Data Level framework
AI Breakdown — "Meta Acquires Moltbook: Facebook for AI Bots"
Runtime: 10 min | Host: Jaden Schaefer | Guest: Matt Schlick, Ben Parr, Andrew Bosworth, Ian All
For the tech investor or executive trying to understand the future of social networks in an agentic world: Meta's acquisition of a "Facebook for AI agents" signals a future where agent-to-agent communication, despite data integrity and security flaws, is the new frontier.
Jaden Schaefer covers Meta's acquisition of Multbook, a social media platform for AI agents. Despite Multbook's controversial history of security vulnerabilities and data integrity issues, Meta's move signals a strategic focus on understanding and orchestrating agent-to-agent communication for their Super Intelligence Lab. This indicates a future where AI agents interact not just with humans, but with each other, setting the stage for a new core layer of the digital world, and potentially validating "dead internet theory" concerns about AI-generated content.
"Every credential that was in Multiple Book Supabase was unsecure for some time. For a little bit of time you could grab any token you wanted and pretend to be another agent on there because it was all public and available." — Ian All
Connects to: AI Agent Communication and Collaboration, Dead Internet Theory, Security vulnerabilities in AI social platforms
Decoder with Nilay Patel — "Anthropic doesn't trust the Pentagon, and neither should you"
Runtime: 49 min | Host: Nilay Patel | Guest: Mike Masnick, Helen Havlik
For the leader navigating complex ethical landscapes in defense contracts and data privacy: This deep dive into Anthropic's legal battle with the Pentagon reveals how government surveillance overreach and reinterpretation of laws fundamentally reshape the ethical considerations of AI deployment.
Nilay Patel and Mike Masnick dissect Anthropic's legal showdown with the Pentagon over AI surveillance, exposing a deeper history of government overreach and the NSA's manipulation of surveillance laws. The core issue is the Fourth Amendment and the "third-party doctrine," which renders much of our digital data vulnerable to seizure without a warrant. Anthropic's firm stance against mass surveillance with AI contrasts sharply with OpenAI's initial "all lawful uses" approach, forcing a critical examination of ethical red lines and the lack of adversarial checks in intelligence operations like the FISA court. This isn't just about AI; it's about the erosion of digital privacy in a hyper-connected world.
"all lawful uses is too big" — Neelai Patel
Connects to: AI ethics and governance, Fourth Amendment and cloud data, Digital privacy in the U.S.
Eye On A.I. — "#325 Phelim Brady: Why AI's Future Depends on Human Judgement"
Runtime: 47 min | Host: Craig S. Smith | Guest: Phelim Bradley
For the CEO looking to future-proof their AI strategy: Understanding why human judgment, not just algorithms, remains the "alpha" in advanced AI development, and how to harness crowdsourced intelligence for superior model performance and ethical deployment.
Phelim Bradley, co-founder and CEO of Prolific, uncovers the "dirty secret" of AI: its deep reliance on human effort for evaluation and refinement. Prolific, a human data platform, provides high-quality, demographically representative human evaluation that challenges traditional academic benchmarks. Bradley argues that human judgment is indispensable for achieving "alpha" in AI—pushing beyond existing capabilities—and ensuring trust and safety, especially in regulated sectors. He highlights the surprising impact of human demographics on model rankings and Prolific's vision to become a full-stack human data platform, even exploring virtual environments for embodied AI data collection.
"I think the human judgment is where the alpha is. So if you're looking to push the capability beyond what is already available, almost by definition you are not able to build an automated evaluator that is able to assess that gap in capability. So you need human judgment." — Phelim Brady
Connects to: Human judgment in AI development and evaluation, Ethical sourcing of data for AI models, AI companies relying on real human feedback for model performance
Last Week in AI — "#236 - GPT 5.4, Gemini 3.1 Flash Lite, Supply Chain Risk"
Runtime: 89 min | Host: Andrey Kurenkov, Jeremie Harris | Guest: Andrey Kurenkov, Jeremie Harris
For the tech lead or investor tracking the frontier of AI capabilities and market dynamics: Get the unfiltered breakdown on OpenAI’s GPT-5.4 Pro release, Google’s Gemini 3.1 Flash Lite, a $110 billion funding round hinged on AGI, and the emerging supply chain risks and ethical battles shaping enterprise AI adoption.
Andrey Kurenkov and Jeremie Harris unpack the latest AI model advancements, spotlighting OpenAI’s GPT-5.4 Pro with its 1M-token context window and Google’s Gemini 3.1 Flash Lite. The conversation delves into the rapid acceleration of AI capabilities, exemplified by Luma AI generating localized ad campaigns for minimal cost. However, it also exposes the growing ethical battleground, specifically the controversy around Anthropic and OpenAI's engagements with the Department of Defense, OpenAI's record $110 billion funding round (contingent on AGI or IPO), and the urgent risks of unaligned AI agents, including real-world manipulation and "mini-catastrophes."
"This incrementation is potentially a symptom of the singularity. I mean, it's actually not wild to suggest that. It could well be hard to know. But in any case, they are coming at us harder and faster than they were before, for sure." — Jeremie Harris
Connects to: AI ethics and governance, OpenAI $110 billion private funding round and $730 billion valuation, AI models causing mini catastrophes by default during training
Hard Fork — "A.I. Goes to War + Is ‘A.I. Brain Fry’ Real? + How Grammarly Stole Casey’s Identity"
Runtime: 67 min | Host: Kevin Roose, Casey Newton | Guest: Julie Bedard
For the leader navigating the human and geopolitical impacts of AI, from employee well-being to military strategy: This episode connects the dots between a new "AI brain fry" phenomenon, AI’s role in modern warfare, and the increasing ethical mess of AI companies co-opting identities.
This episode connects AI's expanding footprint across human and geopolitical domains. It starts with the unsettling reality of AI in modern warfare, as the US and Israel leverage AI tools like Claude for real-time battle planning in conflicts with Iran, making AI infrastructure like data centers strategic targets. It then pivots to the human cost: the emerging phenomenon of "AI brain fry," a new cognitive strain impacting 14% of AI users, especially marketing professionals. Finally, it highlights the ethical quagmire of AI companies co-opting identities, using Grammarly's "Expert Review" feature as a case study, which attributed AI-generated advice to real journalists without consent, sparking a class-action lawsuit. This paints a vivid picture of AI's complex impact on work, ethics, and global conflict.
"AI brain fry... is mental fatigue from excessive use or oversight of AI tools beyond one's cognitive capacity" — Julie Bedard
Connects to: AI in military applications in Iran, AI brain fry, AI attribution and compensation for intellectual property
More from Transformation Brief: AI & Technology
- Episode Guide: The Zero Human Company: OpenAI’s $200M Win and the 80% Inference Wall
- Episode Guide: Pentagon AI Deal Sparks 'Security Theater' Fears
- Episode Guide: Xbox’s AI CEO Bet: 4% GitHub Code by Claude
- Episode Guide: Ring’s “Crime-Free” Dream vs. OpenClaw’s 150 Features/Week
- Episode Guide: Vibe-Coded to $1B: OpenAI Buys OpenClaw as Google Ships 150 Features a Week
Get the next edition delivered to your inbox
