scripod.com
An AI state of the union: We’ve passed the inflection point, dark factories are coming, and automation timelines | Simon Willison

Overview

Shownote

Highlights

Transcript

Chapters

Pins

An AI state of the union: We’ve passed the inflection point, dark factories are coming, and automation timelines | Simon Willison

Lenny's Podcast: Product | Career | Growth

Apr 02
Lenny's Podcast: Product | Career | Growth

Lenny's Podcast: Product | Career | Growth

Apr 02
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Simon Willison is a prolific independent software developer, a blogger, and one of the most visible and trusted voices on the impact AI is having on builders. He co-created Django, the web framework that powers Instagram, Pinterest, and tens of thousands of other websites. He coined the term “prompt injection,” popularized the terms “AI slop” and “agentic engineering,” and has built over 100 open source projects, including Datasette, a data analysis tool used by investigative journalists worldwide. What makes Simon unique is that he’s made the leap from traditional software engineering to AI-native development more fully and visibly than almost anyone—and he’s been documenting everything he learns in real time on his blog, SimonWillison.net [http://SimonWillison.net]. In our in-depth conversation, Simon shares: 1. Why November 2025 was the inflection point when AI coding agents crossed from “mostly works” to “actually works” 2. How Simon writes 95% of his code from his phone now and why he’s mentally exhausted by 11 a.m. 3. Why mid-career engineers (not juniors) are most at risk right now 4. The three agentic engineering patterns Simon uses daily (red/green TDD, templates, hoarding) 5. The next leap: the “dark factory” pattern where nobody writes or reviews code and AI does its own QA 6. Why prompt injection is an unsolved security problem and the “lethal trifecta” that will likely lead to an AI Challenger disaster 7. Why the pelican riding a bicycle became the unofficial benchmark for AI model quality — Brought to you by: WorkOS [https://workos.com/lenny]—Modern identity platform for B2B SaaS, free up to 1 million MAUs Vanta [https://vanta.com/lenny]—automate compliance, manage risk, and accelerate trust with AI — Episode transcript: https://www.lennysnewsletter.com/p/an-ai-state-of-the-union [https://www.lennysnewsletter.com/p/an-ai-state-of-the-union] — Archive of all Lenny's Podcast transcripts: https://www.dropbox.com/scl/fo/yxi4s2w998p1gvtpu4193/AMdNPR8AOw0lMklwtnC0TrQ?rlkey=j06x0nipoti519e0xgm23zsn9&st=ahz0fj11&dl=0 [https://www.dropbox.com/scl/fo/yxi4s2w998p1gvtpu4193/AMdNPR8AOw0lMklwtnC0TrQ?rlkey=j06x0nipoti519e0xgm23zsn9&st=ahz0fj11&dl=0] — Where to find Simon Willison: • X: https://x.com/simonw [https://x.com/simonw] • LinkedIn: https://www.linkedin.com/in/simonwillison [https://www.linkedin.com/in/simonwillison] • Website: https://simonwillison.net [https://simonwillison.net] • Agentic Engineering Patterns: https://simonwillison.net/guides/agentic-engineering-patterns [https://simonwillison.net/guides/agentic-engineering-patterns] — Where to find Lenny: • Newsletter: https://www.lennysnewsletter.com [https://www.lennysnewsletter.com] • X: https://twitter.com/lennysan [https://twitter.com/lennysan] • LinkedIn: https://www.linkedin.com/in/lennyrachitsky/ [https://www.linkedin.com/in/lennyrachitsky/] — In this episode, we cover: (00:00) Introduction to Simon Willison (02:40) The November 2025 inflection point (08:01) What’s possible now with AI coding (10:42) Vibe coding vs. agentic engineering (13:57) The dark-factory pattern (20:41) Where bottlenecks have shifted (23:36) Where human brains will continue to be valuable (25:32) Defending of software engineers (29:12) Why experienced engineers get better results (30:48) Advice for avoiding the permanent underclass (33:52) Leaning into AI to amplify your skills (35:12) Why Simon says he’s working harder than ever (37:23) The market for pre-2022 human-written code (40:01) Prediction: 50% of engineers writing 95% AI code by the end of 2026 (44:34) The impact of cheap code (48:27) Simon’s AI stack (54:08) Using AI for research (55:12) The pelican-riding-a-bicycle benchmark (59:01) The inherent ridiculousness of AI (1:00:52) Hoarding things you know how to do (1:08:21) Red/green TDD pattern for better AI code (1:14:43) Starting projects with good templates (1:16:31) The lethal trifecta and prompt injection (1:21:53) Why 97% effectiveness is a failing grade (1:25:19) The normalization of deviance (1:28:32) OpenClaw: the security nightmare everyone is looking past (1:34:22) What’s next for Simon (1:36:47) Zero-deliverable consulting (1:38:05) Good news about Kakapo parrots — References: https://www.lennysnewsletter.com/p/an-ai-state-of-the-union [https://www.lennysnewsletter.com/p/an-ai-state-of-the-union] — Production and marketing by https://penname.co/ [https://penname.co/]. For inquiries about sponsoring the podcast, email podcast@lennyrachitsky.com [podcast@lennyrachitsky.com]. — Lenny may be an investor in the companies discussed. To hear more, visit www.lennysnewsletter.com [https://www.lennysnewsletter.com?utm_medium=podcast&utm_campaign=show-notes-no-free-preview-language]

Highlights

In this wide-ranging conversation, Simon Willison—Django co-creator and pioneering AI-native developer—offers a candid, deeply practical assessment of how AI is reshaping software engineering, not as a distant future but as an accelerating present.
00:00
AI-pilled people seem to work harder
02:43
GPT-5.1 and Claude Opus 4.5 crossed a threshold in November 2023, making coding agents more reliable and useful for software engineers
08:03
AI enables 'vibe coding' for rapid prototyping but poses risks in responsible use
13:17
StrongDM is conducting experiments on the 'dark factory' pattern for fully automated software production
18:59
Anthropic discovered 100 potential vulnerabilities in Firefox and responsibly reported them
23:01
Real human usability testing delivers more reliable insights than AI-simulated user feedback
23:36
The human brain becomes more valuable when combining and refining ideas
27:52
Their 25-year experience in estimating project time is no longer applicable as AI can complete tasks much faster
29:12
Mid-career engineers may be in trouble as they lack the expertise to use AI effectively and don't get the same benefits as beginners
33:25
AI can never have agency due to lack of human motivations
33:53
Companies let people go due to lack of creativity or ambition
35:12
Those most involved with AI are working harder despite its promise of increased productivity
37:23
Data labeling companies are paying for pre-2022 handwritten code for model training
42:43
The tech job market has the highest number of open engineering and PM roles in 3.5 years globally, excluding the COVID peak.
44:34
Writing code is now much faster, changing the way software engineers work
53:02
Anthropic became the number one app in the app store after enabling seamless memory transfer from ChatGPT to Claude
54:09
Claude is preferred for code and research due to strong search integration
55:14
There's a strong correlation between a model's ability to draw a pelican on a bike and its overall performance.
59:01
ChatGPT fails to draw a pelican on a bicycle
1:05:52
Coding agents can search through large amounts of data to find relevant examples, making them very powerful for coding tasks
1:13:27
AI agents are good at writing tests, making the process easier and less boring
1:14:43
A single test like 1+1=2 serves as an effective thin template to guide AI coding agents' style and behavior
1:19:30
The lethal trifecta requires all three conditions: private data access, malicious instruction exposure, and data exfiltration
1:21:53
Achieving 97% effectiveness against prompt injection is a failing grade
1:25:19
Prompt injection and jailbreaks remain unsolved AI security challenges due to the fuzzy nature of text instructions
1:28:40
OpenClaw went from first line of code in November 2023 to a Super Bowl ad in three and a half months
1:34:24
His goal this year is for his software to contribute to a Pulitzer-Prize winning report
1:36:47
Simon spends an hour on calls with clients without writing reports or code
1:38:06
Dozens of new Kakapo chicks have been born—their first successful breeding season in four years

Chapters

Introduction to Simon Willison
00:00
The November 2025 inflection point
02:40
What’s possible now with AI coding
08:01
Vibe coding vs. agentic engineering
10:42
The dark-factory pattern
13:57
Where bottlenecks have shifted
20:41
Where human brains will continue to be valuable
23:36
Defending of software engineers
25:32
Why experienced engineers get better results
29:12
Advice for avoiding the permanent underclass
30:48
Leaning into AI to amplify your skills
33:52
Why Simon says he’s working harder than ever
35:12
The market for pre-2022 human-written code
37:23
Prediction: 50% of engineers writing 95% AI code by the end of 2026
40:01
The impact of cheap code
44:34
Simon’s AI stack
48:27
Using AI for research
54:08
The pelican-riding-a-bicycle benchmark
55:12
The inherent ridiculousness of AI
59:01
Hoarding things you know how to do
1:00:52
Red/green TDD pattern for better AI code
1:08:21
Starting projects with good templates
1:14:43
The lethal trifecta and prompt injection
1:16:31
Why 97% effectiveness is a failing grade
1:21:53
The normalization of deviance
1:25:19
OpenClaw: the security nightmare everyone is looking past
1:28:32
What’s next for Simon
1:34:22
Zero-deliverable consulting
1:36:47
Good news about Kakapo parrots
1:38:05

Transcript

Simon Willison: A lot of people woke up in January and February and started realizing, oh, wow, I can churn out 10,000 lines of code in a day. It used to be you'd ask ChatGPT for some code and it would spit out some code. And you have to run it and test it...