Artificial Ignorance
Charlie Guo · Writing · Leaning
Artificial Ignorance is Charlie Guo's AI engineering newsletter, with essays, model analysis, and paid tutorials that include code for people building with AI.
About Artificial Ignorance
What Is Artificial Ignorance
Artificial Ignorance is a Substack publication about building with AI, written by Charlie Guo and read by more than 27,000 subscribers. It launched in February 2023 as a way for one engineer to think in public about a field that kept changing under him. Three years later it runs on AI deep dives, model writeups, and hands-on engineering guides.
The focus is narrow on purpose. It's an AI newsletter for developers, product builders, and founders who want to ship something. There's no machine learning theory, no dump of forty links, and no claim that you're already behind. A typical piece answers one question well: how does this model, tool, or pattern behave once it's inside a real product?
The limits matter as much as the strengths. It's text on Substack, so there's no sandbox, no video course, and no forum beyond the subscriber chat. In January 2026 Guo joined OpenAI's developer experience team. He stopped the news roundups at the same time and said he won't share inside information or amplify leaks. His coverage of OpenAI's competitors is still worth reading, with that job in mind.
Getting Started
- Go to ignorance.ai and enter an email address. Free signup takes a few seconds and doesn't ask for a card.
- Open the archive and pick the posts closest to your work: agent design, evaluation, prompting, or model selection.
- Read the free essays for a few weeks. Most deep dives run 1,500 to 3,000 words and state the argument early.
- Upgrade to paid at $10/mo or $100/yr if you want the tutorial track, code samples, and case studies.
- Join the subscriber chat for office hours threads, where paid readers ask about architecture, evaluation, and agent design.
Product Information
A quick look at Artificial Ignorance's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Software engineers building with LLM APIs
- Startup founders and product builders
- Technical leads who need to brief a team
Tasks
- Understanding agent architecture
- Reading a model release past the benchmark table
- Learning the craft layer with code
Scenarios
- A Sunday planning session before a sprint
- Choosing between two models for a feature
- Onboarding a new engineer to AI work
Key features
Engineering Deep Dives
Most issues take a single subject and stay with it for 2,000 words or more. "Agentic Primitives 101" spends its whole length on the harnesses that let a model hold a job across days and conversations. You finish with a working mental model and a list of tradeoffs, not a summary of the news.
Agent and Harness Coverage
Agent design gets more space than anything else. Posts cover tool use, memory, evaluation loops, and the difference between skills, tools, and MCP servers. The voice agent piece argues that a voice model doesn't have to talk back, then lays out three response modes with examples. That kind of framing is hard to find in vendor documentation.
Model Analysis Beyond Benchmarks
New releases get read rather than repeated. Guo goes through system cards, pricing changes, and the gap between a demo and a product, and he says when a model is misunderstood. The GPT-5.3-Codex and Claude Opus 4.6 comparison is a good example: two releases, one post, and a clear view of what each one changes for builders.
Paid Tutorials With Code
The paid tier is where the AI tutorials live. Pieces on retrieval, guardrails, and evaluation come with code you can run, plus notes on the parts that broke. They're written as project walkthroughs rather than documentation, so you see the decisions and the dead ends.
Architecture Case Studies
Case studies describe systems that shipped, including the tradeoffs their teams made. You get the shape of the architecture, the reason behind a queue or a cache, and the cost of getting it wrong. Subscribers also get access to AI fundraising data, which is useful context when you're sizing a market.
Subscriber Chat and Office Hours
The chat is the replacement for the old comment threads. Paid readers post questions about architecture, evaluation, and product integration, and Guo runs recurring office hours in the same space. Topic threads cover stories that don't deserve a full post.
Pros and cons
Pros
- The writer ships software: two decades of programming, a Stanford CS degree, and startup experience show up in the questions he asks about a model.
- Plain language without the hype: a 2,000 word post explains a harness or an evaluation loop without assuming a research background.
- Deep archive: close to 300 posts since 2023, so most problems you're working through already have a relevant piece.
- Cheap next to a course: $10/mo or $100/yr buys the tutorial track for less than most single-day workshops.
- Generous free tier: essays and model writeups stay free, which is enough to decide whether the paid material is worth it.
Cons
- Publishing slowed in 2026: the weekly news roundups are gone, so issues arrive less often than they did in 2024 and 2025.
- Code-heavy material sits behind the paywall: free readers get analysis, while the runnable examples and case studies need a paid plan.
- Text only: there's no video course, no notebook environment, and no way to run an example inside the newsletter.
- A conflict to keep in mind: the author works at OpenAI, so coverage of OpenAI and its rivals deserves a second source before you act on it.
Frequently asked questions
Artificial Ignorance is a subscription newsletter about AI engineering, published on Substack by Charlie Guo. It covers agent design, model releases, and the practical side of building products with LLMs.
Related content
Explore related tools, skills, and articles for Artificial Ignorance.
Artificial Ignorance Alternatives
Emailwritr
Emailwritr · Writing · MarketingEmailwritr is an AI email copywriting tool built for affiliates, course creators, and online businesses that want more revenue from the list they already own. Instead of blank pages or generic chatbot prompts, this AI email writer generates welcome sequences, launch campaigns, cart recovery emails, and promo series from frameworks modeled on well-known direct-response copywriters. You connect your autoresponder, answer a few questions about your offer, and it writes the copy for you.
Swms
Swms AI · Writing · Productivity · BusinessSwms is an AI safety compliance tool that turns a plain description of a job into ready-to-use safety documents. Describe your project, trade, and known hazards, and it writes a safe work method statement, job hazard analysis, safe work procedure, RAMS, or safety data sheet in seconds. It also ships with Oscar, a chat assistant that answers workplace safety questions in any language. Teams in construction, mining, transport, and warehousing use it to cut the paperwork that usually eats into a workday.
Prosp
Prosp · Writing · Voice & Language · MarketingProsp is an AI LinkedIn outreach tool built for agencies and sales teams, and it writes the message and the voice note in your own voice for each prospect so they actually reply. You connect your accounts, find leads, and let the AI draft and send personalized messages at scale, all from one inbox. It's built for people running outbound at volume. That's the whole pitch. Every touchpoint still has to feel human.