How to Trust AI Agents: Verify the Work, Not the Model

How to Use AI on Work You Can't Upload - Offline & Local

For deeper playbooks and analysis: https://natesnewsletter.substack.com/ [https://natesnewsletter.substack.com/] Clean sensitive documents locally: https://unlock-ai.natebjones.com/guides/clean-sensitive-docs-locally [https://unlock-ai.natebjones.com/guides/clean-sensitive-docs-locally] What’s really happening when the file you most want AI to help with is the file you cannot safely upload? The common story is that sensitive work has to stay manual — but the reality is that downloaded models, controlled enterprise systems, and narrow specialist workflows now create several practical paths between “send it to a chatbot” and “do not use AI.” In this episode, I share the inside scoop on how Bayer and Discovery Bank are building private AI specialists, then demonstrates the small version with LM Studio and a synthetic contract on a laptop with the network disconnected. * Why model instructions are not the same thing as a secure product boundary * How a local sensitivity router can flag, mask, and route potentially private material * What LoRA changes when a company tunes a specialist for one narrow job * Where laptop-scale processing ends and managed infrastructure begins * Why open weights do not automatically eliminate platform dependence This matters for operators, builders, security teams, and executives who need useful AI without losing control of confidential files or the learning loop created around them. Subscribe for daily AI strategy and news. Hosted on Acast. See acast.com/privacy for more information. ---------------------------------------- Hosted on Acast. See acast.com/privacy [https://acast.com/privacy] for more information.

I asked Fable and Codex what to automate. They disagreed.

For deeper playbooks and analysis: https://natesnewsletter.substack.com/p/let-ai-pick-what-to-automate [https://natesnewsletter.substack.com/p/let-ai-pick-what-to-automate] What's really happening when you stop telling an AI what to automate and ask it to discover the problem itself? The common story is that AI agents need a tightly specified task — but the reality is that the strongest systems can inspect real work, identify recurring friction, and propose different high-leverage automations. In this video, I share the inside scoop on giving Fable and Codex the same open brief and getting two very different answers. * Why picking the problem is becoming part of the agent's job * How Fable found a strategic editorial preflight opportunity * What Codex built to validate completed content handoffs * Where human judgment still matters * How to turn the method into a reusable automation-discovery skill * For operators, builders, and leaders, the shift is from asking which tool to use to asking which recurring problem is worth solving completely. Subscribe for daily AI strategy and news. Hosted on Acast. See acast.com/privacy for more information. ---------------------------------------- Hosted on Acast. See acast.com/privacy [https://acast.com/privacy] for more information.

The AI Harness Audit: Clean Your Setup Before You Upgrade

Every time an AI missed something, I added another rule. Eventually, the accumulated skills, memories, system prompts, checks, and permissions became a hidden system of their own—and that system was getting in the models' way. In this episode, I audit the harness around my AI and compare what happens when Fable 5 and ChatGPT 5.6 meet compact versus overloaded instruction systems. The audit found 66 skill routes, 172 instruction assets, repeated governance rules, and a discovery layer far beyond Codex's stated budget. The lesson is not simply to shorten every prompt. It is to give each surviving instruction one owner and one reason, load specialist context when the work needs it, and enforce deterministic requirements with hard checks. Privacy Policy: https://www.acast.com/privacy ---------------------------------------- Hosted on Acast. See acast.com/privacy [https://acast.com/privacy] for more information.

Pick an AI Model That Fits How You Actually Work

For deeper playbooks and analysis: https://natesnewsletter.substack.com/ [https://natesnewsletter.substack.com/] What's really happening as the model race expands into GPT-5.6, Fable 5, Grok 4.5, GLM 5.2, and increasingly complicated model mixes? The common story is that you should pick whichever model tops the latest benchmark — but the reality is that the best model depends on how you think, how you prompt, and what your hardest work requires. In this episode, Nate shares the inside scoop on choosing a model by work pattern rather than hype. * Why “dumber” does not mean dumb * How model families develop different working styles * Why benchmarks are evidence, not the selection heuristic * How Ringer pairs a strong architect with cheaper workers * What knowledge-work AI still needs beyond coding harnesses For builders, operators, researchers, and team leaders, understanding your own work is becoming more durable than memorizing every model leaderboard. Subscribe for daily AI strategy and news. Hosted on Acast. See acast.com/privacy for more information. ---------------------------------------- Hosted on Acast. See acast.com/privacy [https://acast.com/privacy] for more information.

AI-Native Companies Run on Code: 15 Rules for Operators

AI made it cheap to build almost anything. Most companies still ship at the old pace, and it isn't because their AI is worse than Anthropic's or OpenAI's. The real difference is what they've moved out of meetings and documents and into working code. Full post: https://natesnewsletter.substack.com/p/ai-native-company-rules?r=1z4sm5&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true [https://natesnewsletter.substack.com/p/ai-native-company-rules?r=1z4sm5&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true] My Links 🔗 👉🏻 Newsletter: https://natesnewsletter.substack.com/ [https://natesnewsletter.substack.com/] 👉🏻 X: https://x.com/natebjones [https://x.com/natebjones] 👉🏻 TikTok: https://www.tiktok.com/@nate.b.jones [https://www.tiktok.com/@nate.b.jones] 👉🏻 Instagram: https://www.instagram.com/nate.b.jones [https://www.instagram.com/nate.b.jones] What's really happening inside the companies that ship AI features every week? The common story is that they just have better AI. The real question is what they rebuilt underneath the model to make that speed possible. In this video, I share the inside scoop on the 15 rules I use to move a company from tool adoption to real AI-native speed: * Why moving repeatable coordination into code is now an operator's job * How 15 commandments work together as one operating system * What changes first: roadmaps, meetings, documentation, and design * Where partial adoption fails, and why every rule has to move together Moving this fast is real, but only if you treat these as one system instead of picking off the rule that feels easiest to adopt. ---------------------------------------- Hosted on Acast. See acast.com/privacy [https://acast.com/privacy] for more information.