Modern Web

Rizel Scarlett, principal developer advocate at Entire and former Goose maintainer, joins Rob Ocel to explore the missing context layer in AI-assisted software development. They unpack why session history, prompts, tool calls, and decision-making need to stay connected to code; how Entire captures that context through Git-based checkpoints; and why AI should be used as a learning partner, not just a code generator. The conversation also covers code-review overload, agent-friendly workflows, and a more sustainable path to adopting AI.

5 Key Takeaways
- AI-generated code makes context and intent more valuable: teams need a clear record of the decisions and experiments that led to a change.
- Entire uses agent hooks and Git-based checkpoints to capture prompts, tool activity, and session context without cluttering the primary codebase.
- AI can be a powerful learning tool: it gives developers a judgment-free way to ask questions, understand unfamiliar code, and build deeper technical intuition.
- Traditional PR descriptions and diff-only reviews struggle with agent-produced changes; reviews need to connect customer pain, intent, design, and implementation.
- Sustainable AI adoption beats chasing every trend: focused, incremental improvements can deliver real value without the cost and stress of full agent autonomy.

Chapters
00:00 — Why AI code needs durable context
04:12 — The documentation and code-review problem
10:11 — AI as a learning tool
13:42 — How Entire captures agent sessions through Git
26:14 — Session handoffs, reusable skills, and better reviews
32:39 — The future of code forges and sustainable AI adoption
41:37 — Where to find Rizel and Entire

Rob Ocel on Linkedin: https://www.linkedin.com/in/robocel/
Rizel Scarlett on Linkedin: https://www.linkedin.com/in/rizel-bobb-semple/
This Dot Labs Twitter: https://x.com/ThisDotLabs
This Dot Media Twitter: https://x.com/ThisDotMedia
This Dot Labs Instagram: https://www.instagram.com/thisdotlabs/
This Dot Labs Facebook: https://www.facebook.com/thisdot/

Sponsored by This Dot: https://www.thisdot.co/

AI models are becoming more capable, but reliability remains one of the biggest challenges. In this episode of the Modern Web Podcast, Rob Ocel and Brandon Mathis are joined by Andrew Baker, Founder & CEO of Balon.ai, and Arif Hosein, COO of Balon.ai, to discuss why AI hallucinations persist and what it will take to reduce them.
The conversation explores Balon.ai's Deep Fusion architecture, which combines multiple AI models through a consensus-driven approach to improve accuracy without sacrificing performance. They also discuss why larger models alone won't solve reliability, the tradeoffs between cost and quality, and what engineering teams should consider as AI moves deeper into production systems.

Chapters
00:00 Introduction & Meet the Guests
01:20 The AI Hallucination Problem
04:32 What Is an AI Hallucination?
08:22 Sponsor Message
08:47 Are Hallucinations Getting Better?
12:53 Balancing Accuracy, Agency & Emergent Behavior
14:07 Deep Fusion: Balin AI's Multi-Model Architecture
17:20 How Consensus Works Across Multiple Models
20:20 Simultaneous Peer Review for AI
22:34 Why Multiple Imperfect Models Can Produce Better Results
27:58 Can Deep Fusion Fix Bad Prompts?
30:41 The Cost of Multi-Model AI
34:21 AI ROI, Vibe Coding & Enterprise Reality
37:56 How to Try Balin AI
39:37 Final Thoughts & Outro

Rob Ocel on Linkedin: https://www.linkedin.com/in/robocel/
Brandon Mathis on Linkedin: https://www.linkedin.com/in/mathisbrandon/
Andrew Baker on Linkedin: https://www.linkedin.com/in/andrewbakeratl/
Arif Hosein on Linkedin: https://www.linkedin.com/in/arif-hosein/
This Dot Labs Twitter: https://x.com/ThisDotLabs
This Dot Media Twitter: https://x.com/ThisDotMedia
This Dot Labs Instagram: https://www.instagram.com/thisdotlabs/
This Dot Labs Facebook: https://www.facebook.com/thisdot/

Sponsored by This Dot Labs: https://www.thisdot.co/

Rob Ocel sits down with Minko Gechev to unpack one of the fastest-evolving areas in AI-assisted development: skills, agent workflows, and the growing complexity of AI harnesses.Explore AI skills, how they differ from rules and MCPs, and why developers are increasingly treating skills as structured, executable instructions rather than traditional documentation. Minko explains how modern agents operate under the hood, how skills are lazy-loaded into context windows, and why poorly designed skills can create context rot, unnecessary token usage, and degraded performance over time.
Rob and Minko also dig into one of the most important but misunderstood topics in AI engineering today: evals. They break down how teams can evaluate AI workflows, measure pass rates, test deterministic versus non-deterministic outputs, and optimize skills for reliability, token efficiency, and execution quality across evolving models and harnesses. Also learn about prompt structure, workflow orchestration, context compression, state-machine agents, React loops, and why many teams are accumulating rules and skills faster than they are pruning them.

What you will learn:
• Learn how engineering teams are moving AI beyond isolated experiments into repeatable workflows
• Explore strategies for improving reliability, testing, and validation in AI-assisted development
• Hear practical approaches to scaling AI adoption across engineering organizations
• Understand how leaders are thinking about automation, CI/CD, and operational maturity in the age of AI
• Connect with technology leaders discussing the challenges and opportunities shaping modern software delivery

Rob Ocel on Linkedin: https://www.linkedin.com/in/robocel/
Minko Gechev on Linkedin: https://www.linkedin.com/in/mgechev/
This Dot Labs Twitter: https://x.com/ThisDotLabs
This Dot Media Twitter: https://x.com/ThisDotMedia
This Dot Labs Instagram: https://www.instagram.com/thisdotlabs/
This Dot Labs Facebook: https://www.facebook.com/thisdot/

Sponsored by This Dot Labs: https://www.thisdot.co/

In this episode of the Modern Web Podcast, Rob Ocel talks with PlanetScale CEO Sam Lambert about what “database scale” actually looks like in 2026. Sam shares migration stories from companies that moved at the perfect time and others that waited until they were already in trouble, plus why sharding and reliability are never just “magic” if your queries and data model are a mess.
They also cover PlanetScale’s evolution beyond its MySQL and Vitess roots into Postgres, Metal, and what’s coming next for sharding in the Postgres world. Along the way, they connect the dots to AI workloads, which are increasingly write heavy and put new pressure on performance, uptime, and security.

What You'll Learn:
- How to spot the “right time” to migrate databases before you’re on fire (and what happens when you wait too long)
- What PlanetScale actually gives you “for free” at scale, and what it can’t fix (bad schema, missing indexes, terrible SQL)
- Why “auto” database magic is usually a tradeoff, and what to ask for when you want to peek behind the curtain
- What PlanetScale is becoming beyond MySQL/Vitess, including Postgres, Metal, and Nikky (sharding for Postgres)
- How AI workloads are changing database patterns, especially the shift toward write heavy systems and why that pressures reliability and security

Sam Lambert on Linkedin: https://www.linkedin.com/in/isamlambert/
Rob Ocel on Linkedin: https://www.linkedin.com/in/robocel/
This Dot Labs Twitter: https://x.com/ThisDotLabs
This Dot Media Twitter: https://x.com/ThisDotMedia
This Dot Labs Instagram: https://www.instagram.com/thisdotlabs/
This Dot Labs Facebook: https://www.facebook.com/thisdot/

Sponsored by This Dot Labs: https://www.thisdot.co/

Tracy Lee and A.D. Slaton sit down on The Context Window to unpack a wild week in AI, starting with the eye popping 500 billion dollars spent on AI infrastructure in 2025 and why the Cognizant CEO still says enterprise value is missing. They dig into reports of ChatGPT going “code red” in response to Gemini 3, what that means for OpenAI, and what it means for everyday builders trying to ship real products. Along the way they touch on ByteDance, call out LiveKit as a key piece of infrastructure for voice, video, and physical AI agents, and flag IBM’s move to acquire Confluent as another signal of where data and AI are heading.

What you will learn
- Why 500B spent on AI infrastructure has not translated into clear enterprise value yet
- What the Cognizant CEO’s comments really signal for teams building AI products
- How Gemini 3’s launch is shaking up the landscape for ChatGPT and OpenAI
- What a “Code Red” moment actually means for developers and companies relying on these platforms
- How LiveKit powers voice, video, and physical AI agents and where it fits in the stack
- Why IBM acquiring Confluent matters for data, streaming, and real time AI systems
- How to stay grounded and make practical decisions when AI news makes reality feel unstable
0:00 Intro
0:53 Are we overspending on AI infrastructure and where’s the enterprise value
2:54 Adoption gap, enablement work and why 100% AI generated code is still rare
6:11 High touch AI training, workshops and scaling AI practices across teams
8:58 Grok 4.22, AI trading experiments and quant style tools for everyone
13:51 OpenAI “Code Red,” rising competition and what changes for Agile with agents
20:37 ByteDance agentic phone, AR glasses and AI moving into the physical world
23:20 LiveKit, voice cloning, AI podcasts and the problem of AI slop
27:00 Thinking machines, social media’s role in AI and closing reflections

Tracy Lee on Linkedin: https://www.linkedin.com/in/tracyslee/
A.D. Slaton on Linkedin: https://www.linkedin.com/in/adslaton/
This Dot Labs Twitter: https://x.com/ThisDotLabs
This Dot Media Twitter: https://x.com/ThisDotMedia
This Dot Labs Instagram: https://www.instagram.com/thisdotlabs/
This Dot Labs Facebook: https://www.facebook.com/thisdot/
This Dot Labs Bluesky: https://bsky.app/profile/thisdotlabs.bsky.social
Sponsored by This Dot: https://ai.thisdot.co