Built This Week

DNA is just another language.In Episode 32 of Built This Week, we sit down with Dov Gertz, founder of Converge Bio, to explore how generative AI is transforming drug discovery.Every human can be represented as 3.2 billion nucleotides built from four letters: A, C, G, and T. If computers run on zeros and ones, we run on biological code.Converge Bio is training frontier foundation models on DNA, RNA, proteins, and small molecules — helping biotech and pharma companies design better drugs, faster and cheaper.We also demo a retro-inspired “Cell Defense Arena” game built for Converge to use at conferences.Then we pivot into AI infrastructure and agent workflows:The GPU bottleneck and pharma’s growing demand for compute Why molecular AI is 5 to 10 years behind text models How AI could reduce drug timelines from 10 years to 6 to 8 Why cancer and autoimmune diseases may benefit first The limits of FDA regulation in shortening approval cycles OpenClaw, multi-agent systems, and infinite AI teams Cloud versus on prem in the era of foundation modelsThe big takeaway:Chatbots are impressive. But AI applied to biology could extend human life.If you work in biotech, pharma, AI research, or frontier infrastructure — this episode is for you.New episodes every Friday.⏱ TIMESTAMPS(0:00) DNA as code: 3.2 billion nucleotides (0:32) Welcome to Episode 32 (1:00) Meet Dov Gertz and Converge Bio (2:02) Demo: Cell Defense Arena game (3:25) Converge Bio’s $33M raise and mission (4:05) Foundation models for molecular data (5:00) Turning DNA, RNA, and proteins into machine-readable text (6:02) How transformers apply to biology (7:03) 400x more DNA than text on the internet (8:02) Who Converge’s customers are (9:21) Faster, cheaper, better drug discovery (10:39) The three bottlenecks: data, architecture, compute (12:02) The future of personalized medicine (13:02) Which diseases benefit first: cancer, diabetes, autoimmune (14:00) Regulatory realities and clinical trial timelines (16:30) Will AI shorten drug approval cycles? (17:01) NVIDIA, GPUs, and scaling molecular AI (18:30) Pharma as a new AI infrastructure consumer (19:13) Hard pivot: OpenClaw and agentic AI (21:26) Managing teams of AI agents (22:20) Cloud versus on prem debate (25:02) Why developers must adapt weekly (29:26) Closing thoughts and where to find Converge Bio🔗 LINKSConverge Bio https://converge-bio.comBuilt This Week New episodes every Friday https://builtthisweek.comJordan Metzner https://x.com/mrjmetzSam Nadler https://x.com/Gravino05

Private equity due diligence used to take hundreds of hours. Now it takes seconds.In Episode 31 of Built This Week, we sit down with August Kiles, Head of Product at Emblem, to break down how AI is transforming investment funds — from venture capital to growth equity to private equity.Emblem is building what they call the “last platform investors will ever need” — a system that ingests entire data rooms, extracts financials, compares deals, generates reports in Word, Excel, and PowerPoint, and helps funds get to a “no” faster.We also demo a portfolio scenario simulation tool inspired by Emblem — showing how macro events like regulatory pressure or liquidity surges could impact a 30-company portfolio.Then we dive into the latest AI news:Amazon engineers pushing for Claude Code over internal toolsWhy Opus 4.6 is a step-function improvement for codingHow AI is changing software development workflowsElon Musk’s XAI reorg and what it signals about model competitionThe big takeaway:AI is not eliminating analysts. It’s increasing deal throughput and freeing them to focus on alpha.If you work in VC, private equity, family offices, or growth equity — this episode is for you.New episodes every Friday.⏱ TIMESTAMPS(0:00) Emblem’s mission: the last platform investors will ever need (0:25) Welcome to Episode 31 (0:55) Meet August Kiles from Emblem (1:28) Building a portfolio scenario simulation tool (2:05) Modeling regulatory pressure across a 30-company fund (3:00) Liquidity supernova scenario explained (4:00) What Emblem actually does for investment funds (5:00) AI-powered due diligence and data room indexing (6:00) From 100 hours of analysis to seconds (7:20) The old way vs the AI-powered way (8:30) Will AI reduce analyst headcount? (9:40) Getting to “no” faster in private equity (10:30) Where Emblem shines: seed vs private equity (12:00) Multi-agent model orchestration inside Emblem (13:00) How new models improved financial modeling (15:00) Amazon engineers pushing for Claude Code (17:30) Step-function improvements in Opus 4.6 (19:00) Coding workflows transformed by new models (21:30) Elon Musk’s XAI reorganization (23:00) Why model quality now matters more than IDE (25:00) Final thoughts and wrap-up🔗 LINKSEmblem https://emblem.peBuilt This Week New episodes every Friday https://builtthisweek.comJordan Metzner https://x.com/mrjmetzSam Nadler https://x.com/Gravino05

The biggest shift in AI isn’t a new model. It’s agents managing other agents.In Episode 30 of Built This Week, Sam Nadler and Jordan Metzner break down how they’re actually using the latest AI releases — including Claude Opus 4.6 and OpenAI Codex 5.3 — to build real software inside their own workflows.Jordan walks through a private, fully local AI system built with Claude Code that turns raw 23andMe data, blood work, medications, and personal health inputs into a unified health dashboard. The goal isn’t diagnostics — it’s creating a long-term, living record that surfaces insights doctors don’t easily connect.Sam then demos an AI-powered personal trainer built using the new Codex desktop Mac app and high-reasoning models. The system adapts workouts rep-by-rep, adjusts volume in real time, and highlights the tradeoffs between fast iteration tools and slower, deeper reasoning workflows.We close with the biggest AI platform launches of the week:Anthropic’s Opus 4.6 and Agent TeamsOpenAI Frontier and enterprise AI coworkersPerplexity’s Council Mode and LLM swarmsThe era of one chatbot at a time is over. The new skill is learning how to manage AI agents that manage other agents.No hype. No abstractions. Just what actually happens when builders use AI on themselves first.New episodes every Friday.TIMESTAMPS(0:00) The shift from single-agent AI to multi-agent systems (0:21) Welcome to Built This Week Episode 30 (1:00) Agenda and why this week matters (1:38) Why Jordan downloaded his 23andMe data (2:30) Turning unreadable DNA files into usable insights (3:50) Combining genetics, blood work, and medications (5:05) Drug response insights and hereditary signals (6:10) Generating doctor-ready reports for family (7:20) Why this system runs fully local (8:00) Building personal software instead of buying tools (8:40) Sam’s AI personal trainer built with Codex (9:50) Rep-by-rep workout feedback and fatigue detection (10:45) Designing AI interfaces for real-world use (11:40) Codex vs Claude Code: speed vs deep reasoning (12:20) Anthropic Opus 4.6 and Agent Teams (13:00) OpenAI Frontier and AI coworkers (13:25) Perplexity Council Mode and model swarms (14:05) Why multi-agent management is the real inflection (15:15) Becoming a manager of AI managers (16:00) How many agents one human can manage (17:00) AI’s impact on legacy software companies (18:15) Episode 30 wrap-up and what’s nextLINKSBuilt This Week New episodes every Friday https://builtthisweek.comJordan Metzner https://x.com/mrjmetzSam Nadler https://x.com/Gravino05

Can you really build serious internal AI tools in a few hours — and should everyone on your team be doing it?In Episode 29 of Built This Week, Sam Nadler and Jordan Metzner break down an internal AI product they built at Ryz Labs called ScreenEval — a recruiter screen analysis and coaching tool built in under six hours using Claude Code, Supabase, and AWS.We start with a live demo. Sam walks through how ScreenEval ingests recruiter screen transcripts, evaluates candidates, scores recruiter performance, and provides concrete coaching feedback — all without overriding human judgment. The real unlock is turning messy interview transcripts into searchable, structured hiring data across the entire organization.From there, we test Claude Cowork live — Anthropic’s new interface designed to make building accessible to non-technical users — and compare it to running Claude Code directly in the terminal. We discuss where Cowork shines, where terminal-based workflows still win, and why managing multiple AI agents is becoming a core skill.We wrap with AI news, including Anthropic’s massive funding round, pricing changes, and why enterprise-focused AI tooling is pulling spend away from other platforms.No hype. No abstractions. Just what actually happens when you put AI to work inside a real company.New episodes every Friday.================================================================================TIMESTAMPS(0:00) Why internal AI tools matter more than external products (0:55) Episode 29 kickoff and overview (1:45) Why Ryz Labs built ScreenEval (3:30) Live demo: recruiter screen transcript analysis (6:15) Candidate evaluation vs recruiter coaching (9:10) What recruiters miss in fast screening calls (11:40) AI feedback that doesn’t override human judgment (14:00) Searching transcripts instead of resumes (17:20) Manager dashboards and recruiter performance analytics (21:10) How long it actually took to build ScreenEval (23:30) The full stack: Claude Code, Supabase, AWS (25:45) Why Anthropic models power everything (27:30) Claude Cowork explained (29:15) Building a new product live with Cowork (32:40) Cowork vs Claude Code in the terminal (36:00) Managing multiple AI agents at once (39:30) Anthropic’s funding round and market momentum (42:15) Why we’re shifting spend away from other AI tools (45:10) AI inside organizations: efficiency without layoffs (48:30) What every team should be building next (50:45) Final thoughts and closing================================================================================LINKS SECTIONBuilt This Week New episodes every FridayJordan Metzner https://x.com/mrjmetzSam Nadler https://x.com/Gravino05Built This Week https://builtthisweek.com

Can you really hack an airplane? And if so, how do you test for it without grounding the fleet for a year?In Episode 28 of Built This Week, Sam Nadler and Jordan Metzner sit down with Eero Salih, CTO of Syberian, to explore how AI is transforming cybersecurity for commercial aviation.We start with a live demo — a flight ops cyber radar Sam built to surface real-time security risks across airline operations. Then Eero breaks down what Syberian actually does: building digital twins of aircraft systems to run risk assessments without ever touching the physical plane.This is critical because traditional penetration testing would ground an aircraft for up to a year for recertification. Syberian's AI-powered approach analyzes over 100 technical documents to map every computer system on board — from avionics to entertainment to crew scheduling — and identify vulnerabilities before they become incidents.We also discuss:Why cyber attacks on aviation are now classified as safety threatsNew 2026 regulations forcing airlines to comply with stricter cybersecurity standardsHow small teams are replacing developers with AI agent managersThe tools Syberian uses: Claude Code, Windsurf, Anthropic, and GeminiWhy Google and Anthropic are rejecting ads while OpenAI explores themAn ex-Amazon exec who vibe-coded a full CRM replacement in 72 hoursNo hype.No theory.Just what happens when you put AI in charge of protecting critical infrastructure.New episodes every Friday.================================================================================TIMESTAMPS--------------------------------------------------------------------------------(0:00) Why you can't hack-test an airplane(0:45) Episode 28 kickoff and guest introduction(1:30) Live demo: Flight ops cyber radar dashboard(3:00) Analyzing real-time security threats across airline systems(4:30) What Syberian actually does (in plain English)(6:00) Why physical penetration testing grounds planes for a year(7:30) Using AI to build digital twins of aircraft systems(8:15) Hiring managers, not developers — AI agents do the coding(9:30) Tools of the trade: Claude Code, Windsurf, Anthropic, Gemini(10:00) New 2026 aviation cybersecurity regulations explained(11:00) How cyber attacks became classified as safety threats(12:30) The ripple effects: baggage weight, fuel calculations, pilot tablets(13:30) Who are Syberian's customers? Airlines, private jets, and more(14:55) AI News: Google and Anthropic reject ads in chatbots(16:30) Why Anthropic's no-ads stance matters for enterprise customers(17:30) Amazon exec vibe-codes full CRM replacement in 72 hours(18:30) Why vibe coding works for internal tools but not production(19:15) Final thoughts and closing================================================================================LINKS SECTION--------------------------------------------------------------------------------Built This WeekNew episodes every FridayJordan Metznerhttps://x.com/mrjmetzSam Nadlerhttps://x.com/Gravino05