FeedworthyAI
Technology

Nate’s Substack

natesnewsletter.substack.com

Daily newsletters on AI strategy, news, and implementation for practitioners and leaders who are past the hype and ready to build.

podcastTechnology

Articles89

OpenAI's agent manual rotted into a graveyard of stale rules. Grab my Working Context Starter Kit, the four-file guide that keeps yours current.

Executive Briefing: Your Team Will Believe the Layoff Headline Over Your Roadmap. Here's the Fix.

11,755 agent runs, and the ones that lied looked the most finished. Here are the three checks you can run today (+ my Mission Fit Skill)

Nobody Checked Deloitte's Report. One Academic Did. It Cost Them A$97,587.

Your AI bet can be right and still run out of money. Grab the two-clock prompt: what has to be true, how long it really takes, who controls your runway, what pays today.

Executive Briefing: Which of the 5 Levels of AI Builder Are You, and What It Costs You

Somebody else decided what good looks like, and it shipped with the skill you installed. Here's the guide to fix it.

I Built The Token Saver Skill To Cut My Token Use By 90%. Here Is What It Can And Cannot Do For You.

Stop guessing whether a cheaper model can do the job. Grab the bakeoff guide: the validator, the manifest, the score sheet, and the fixtures.

Executive Briefing: Gumroad Let a Customer Approve Its Code. Here's Where Your Agent Should Stop

Your company blocked ChatGPT for sensitive files. Grab the guide to strip the name, the address, and the price, and the block stops mattering.

Substack co-founder Chris Best on AI slop, detection, and what still counts as thinking.

Kimi K3 is downloadable. That doesn't mean you can run it.

Executive Briefing: How Microsoft, Bayer, and Discovery Use AI on the Data You Can't Upload

I asked Fable and Codex what my business should automate. They disagreed.

I gave Fable 5 five thousand extra words of instructions. It thought better and failed delivery two runs out of three.

Fable 5 is the smarter model. I still open GPT-5.6 Sol every day — Model Fit will tell you which one is yours.

Executive Briefing: Point an agent at your calendar and your repo, and it will show you the rules your company is actually running. Here are the 15 I wrote for mine.
Executive Briefing: Point an agent at your calendar and your repo, and it will show you the rules your company is actually running. Here are the 15 I wrote for mine.

Grab the One-Minute Test That Tells You If Your Task Needs a Chat, One Agent, a Team, or Nothing at All

Stop waiting for AI you can trust. Borrow the 500-year-old trick that made untrustworthy agents useful anyway. (Yes, there's a no-code guide!)

Executive Briefing: Run the $40 question on your org this week. If nobody can answer it, you've found your real AI bottleneck.

Fewer than 1% of denied insurance claims get appealed, and a third to half of appeals win. Build the AI rig that turns your denial and tax pile into cited packets — it drafts, never sends.
Fewer than 1% of denied insurance claims get appealed, and a third to half of appeals win. Build the AI rig that turns your denial and tax pile into cited packets — it drafts, never sends.

Stop paying frontier prices for work a cheaper AI would crush. Grab the model-picker prompt that routes the deck, the repo, and the call.

You can build 80% of your own AI memory by talking to the agent already on your computer

Run this 4-question test before you let any AI into your files, your Slack, or your phone.

Executive Briefing: Cheap Intelligence Won’t Matter If Your Context Is Trapped

Grab the Open Engine guide: the copy-paste task record that makes one AI's work the next AI's job, with receipts

The Five Questions That Turn a Messy Task Into an AI Loop (+ the prompts to map yours)

Grab the 5 prompts that get you ready for Fable 5 before it's back

Executive Briefing: Your team is running agents nobody owns. The one-page card and two prompts that fix it.

Your skills are leaving your hands. Don't let a rent-a-brain keep them.

Vercel deleted 80% of its agent's tools and the agent got better + what to delete from yours (guide inside!)

Executive Briefing: Your company is about to get cheap intelligence. That is not the same as being able to use it.

Grab my Ultimate Guide to Codex and catch up to the 1 in 1,600 people using it every week (mostly no code!)

I heard you
I heard you

Claude vs. Codex isn't about code. It's about whether you steer or dispatch.

Executive Briefing: Uber Burned Its Entire AI Budget Early. The Bill Was Trying to Tell Them Something.

You can't trust one token number across your tools. Here's the guide to a dashboard that keeps Codex, Claude, and ChatGPT honest.

Opus 4.8 scored 81 in my benchmark. I still wouldn't default to it. (The full breakdown + Nate's Community Slack)

Why I’m moving this Substack from daily coverage to deeper weekly work

Executive Briefing: Your career evidence is thinner than you think + 3 prompts that rebuild it

Your prototype graveyard is leaking secrets. The Prototype Classifier + Demotion Audit decide what stays
Your prototype graveyard is leaking secrets. The Prototype Classifier + Demotion Audit decide what stays

Your agent dashboard is green. The run underneath it is where the work actually broke.

The deck got forwarded with a wrong number inside. The Trust Layer's two-model review is built to catch exactly that.

One person's best AI session vanishes the second they close the tab. Grab the 3 prompts that make it your team's.

AI made your app teams 10x faster. Nobody gave your platform team 10x the headcount.

Executive Briefing: Your AI vendor contract isn't built for a capacity crunch. 3 prompts to fix it before your budget meeting

Build the room before you write the memo. Grab the 4-prompt project room kit: source inventory, duplicate log, missing-context list, grounded draft.

68% of AI power users do one thing differently — and it is not a prompt trick
68% of AI power users do one thing differently — and it is not a prompt trick

Seven questions decide whether your AI agent ships. Most teams can answer two.
Seven questions decide whether your AI agent ships. Most teams can answer two.

Six agent protocols just launched. Three of them decide which products survive. Here is how to tell which three.
Six agent protocols just launched. Three of them decide which products survive. Here is how to tell which three.

What ChatGPT sees when it looks at your company + 3 diagnostics

Executive Briefing: Stop asking if AI can do this. Start asking what shape the work is.
Executive Briefing: Stop asking if AI can do this. Start asking what shape the work is.

Exclusive: a conversation with Tibo from Codex on what your company has to become when the model can actually do the work
Exclusive: a conversation with Tibo from Codex on what your company has to become when the model can actually do the work

The 2 prompts I'd run before any 2026 SaaS renewal (especially if you're deploying agents)
The 2 prompts I'd run before any 2026 SaaS renewal (especially if you're deploying agents)

Six things have to be true before AI changes a workflow. Most companies have built two.
Six things have to be true before AI changes a workflow. Most companies have built two.

Your AI agent is rediscovering 85% of its context every run. Here's the architecture fix (+ Contract Spec, Failure Triage, and Stack ADR)

Six layers your agent has to handle. Most products have only thought about two. + a responsibility-layer audit.
Six layers your agent has to handle. Most products have only thought about two. + a responsibility-layer audit.

You gave your AI agent real tools. Here's the 4-part control layer it's missing + the Judge Layer implementation guide

Executive Briefing: Six announcements in 48 hours just changed how enterprise AI gets bought (+ 2 prompts for the new process)

OpenAI made Codex smart enough that the bottleneck moved. Most people haven't noticed where it went.

271 bugs found in Firefox, zero written by a human attacker. What this means for the future of safe code + 2 prompts
271 bugs found in Firefox, zero written by a human attacker. What this means for the future of safe code + 2 prompts

OpenClaw, Anthropic, and Gemma 4 just redefined what "agent framework" means. You need to pick a side.
OpenClaw, Anthropic, and Gemma 4 just redefined what "agent framework" means. You need to pick a side.

The next AI platform winner won't have the best model. They'll own something most companies don't even see yet.

The Anticipation Gap: Why 4 Problems Have to Be Solved Together for Consumer AI to Work

55-75% of your week is on thin ice. Here is the audit that shows you which part.

Executive Briefing: What Stripe Sessions 2026 actually means for how you sell

AI agents are about to route around every tool that can't pass 5 structural tests. Here's the diagnostic.

The buying rule for your personal AI computer (and how to skip the $5,000 mistake)

The four-hour-a-week tax you are paying because IT picked the wrong AI default

The 5-question filter I run every agent launch through (so you can stop reading release notes)

ChatGPT 5.5 scored 87 where the next best model scored 67. Here's what that gap looks like in real work.

Your team spends 5 hours a week on work a sales consultant automated in an afternoon + the 2 prompts that find your version

Executive Briefing: The AI cost curve your strategy is riding just broke + 3 prompts to find your exposure

What GPT-Image-2 actually changed — and the creative ops function that makes you the one who compounds from it

Claude Design just cut 60% of your designer's week — here's what to do with the rest + 4 prompts

Your automation strategy has a blind spot the size of your entire legacy stack. Codex just filled it.

Karpathy's viral AI wiki has a flaw most of the 100K people who bookmarked it haven't noticed yet

Opus 4.7 is smarter, more literal, and quietly more expensive. Those are three different problems.

Your Comprehension Is Worth More Than Your Output Now. Here's How to Make It Visible (Nate's TalentBoard)

Executive Briefing: Why Your World Model Will Look Authoritative for Six Months and Wrong at Year Two

The $300 Overnight Loop That's About To Eat Your Competitive Advantage

The Six-Month AI Context You Lose Every Time You Switch Tools, Jobs, Or Employers

You're Spending Six Figures on AI Models. The Bottleneck Is a 4-Minute CI Pipeline — and Nobody's Fixing the Right Thing.

Your agent needs a SOUL.md you can't write from scratch. I built a 45-minute prompt that writes it for you.

Sora died. Atlassian cut 1,600 engineers. Anthropic got blacklisted. The thread that connects them runs through your org.

Your codebase is full of code nobody understood — not when it shipped, not now, not ever. Here's the fix.

Executive Briefing: Valve Got Lord of the Flies. Zappos Got Paralysis. Your Reorg Is Next.

GPUs Just Got 6x More Valuable. No New Hardware Required.