/_astro/ai_writing_index.Bmx5anHU.webp

AI Writing Index, Feb '23 to Jun '24

My early AI work, in index form, with reflections

JACK IVERS ESSAY 4 MIN READ

As I dig in to recap what I saw and learned at the AI Engineer World’s Fair 2025 in San Francisco, I was forced to face my failure to write consistently – a failure that dates back almost exactly a year, to June 12, 2024. Thinking back on it, I see two causes. First, my book of business on the fractional CTO side of things filled up completely, reducing available writing time. Second, the AI emergence, like Solomon Epstein’s ill-fated space yacht in The Expanse, just kept accelerating. I had “gotta write about that!!” panics daily, until I finally had to just stop rather than driving myself nuts.

I do have to say, though, that during the 16 months from February 2023 to June 2024:

  • I was early to AI
  • I wrote a lot
  • It was pretty decent stuff
  • My opinions and predictions regularly proved accurate. (One prediction, that Apple might get pinched by the AI emergence, is playing out in front of our eyes right now. Apple had little AI-related to announce at this week’s WWDC, and had no visibility whatsoever at AIEWF2025.)

It’s also pretty amazing to see what has happened in the past year. At AIEWF2025, Simon Willison was forced to cut the scope of his LLM update keynote from “the past year” to “the past six months” to have a prayer of finishing in the allotted 20 minutes.

Anyway, take a look at some of the stuff below and see what you think.

P.S.: In June 2025, Cursor and LLMs are vastly better at writing Markdown than a year ago.

Blog Posts

TitleDateDescription
Using Multimodal AI to Capture and Enrich Heirloom Recipes2024-02-19I applied OpenAI’s GPT-4 Vision model and Chat Completions API to preserve a treasure-trove of legacy family recipes.
An Appllama Week in AI2023-07-21Meta makes waves with Llama 2, while Bloomberg pumps itself with Apple LLM non-news.
Really, OpenAI?2023-07-07ChatGPT+ subscribers can prevent OpenAI from using their inputs as training data. That is, so long as they forego the service’s second-best feature.
Of Moats and Moat Busters2023-06-29Open Source, Commercial-friendly AI Challenges the Major Closed AI Players
Unlimited Stupid Questions2023-06-14Possibly the most important benefit of ChatGPT …
Is Runway Gen-2 Midjourney for Video?2023-06-08We Tested Gen-2 and Share the Resulting Videos
The Emergence2023-04-18Season 1, Episode 4
March Madness, AI Style2023-03-18Observations on AI’s March acceleration
Alarmed About AI?2023-02-20John Seely Brown, former director of Xerox PARC, has helpful advice—from April 2000!
Accelerating Crafty’s Launch Using AI Models2023-02-08Thanks to AI-based tools like Midjourney, Crafty shaved at least a month off its time to launch.

Micro Posts

TitleDateDescription
Apple Intelligence Initial Thoughts2024-06-11Mostly from other people, I’m still digesting
A Quick Look at GPT-4o2024-05-19I did a quick benchmark of the new GPT-4o model versus GPT-4-turbo. Roughly twice the speed, with improved Vision accuracy.
OpenAI Vision + Function Calling Examples2024-04-28OpenAI’s April 9th model update merges Turbo and Vision, and these examples show the two working together.
Simon Willison’s LLM Tool: Now I Have 50 LLMs2024-03-25I updated my installation of Willison’s LLM tool to add plugins, and now I have 50 LLMs at my fingertips, including 15 local models, which get installed on demand.
Did GPT-4 Code Interpreter Escape From Its Sandbox?2023-07-14I gave Code Interpreter a workout this morning, and it appeared to exit the building.
ChatGPT+ Privacy Update2023-07-10There appears to be a workaround for now to OpenAI’s “If you want privacy, no chat history for you!” policy.
Llama 22023-07-19HN comments on Llama 2.
Apple Really Did Ignore the AI Emergence at WWDC2023-06-18As I watched the WWDC Keynote and Platforms State of the Union from this year’s WWDC, I was amazed that Apple appeared to be ignoring the current massive emergence in generative AI. Later, I confirmed it.
ChatGPT Code Review2023-06-03Another great use: coming up to speed on an unfamiliar codebase.
Working With My ChatGPT Intern2023-05-29Yesterday was a new record.

Elsewhere

TitleDate
Two Worthwhile Reads From Simon WillisonMarch 25, 2024
Willison: The killer app of Gemini Pro 1.5 is videoFebruary 22, 2024
Ethan Mollick: “How to Use AI to Do Stuff”July 26, 2023
OpenAI Concedes: AI Can’t Detect AIJuly 26, 2023
”Commoditizing the Petaflop” with George Hotz of the tiny corpJuly 9, 2023
Just-released GPT-4 Code Interpreter is a Big Deal, Part 1July 11, 2023
Just-released GPT-4 Code Interpreter is a Big Deal, Part 2July 11, 2023
The Lone Banana ProblemJuly 5, 2023
Midjourney V5.2 (Ars Technica)June 24, 2023
I-JEPA: The first AI model based on Yann LeCun’s vision for more human-like AIJune 19, 2023
Comparing Adobe Firefly, Dalle-2, OpenJourney, Stable Diffusion, and MidjourneyJune 21, 2023
Faster sorting algorithms discovered using deep reinforcement learningJune 7, 2023
It’s infuriatingly hard to understand how closed models train on their inputJune 4, 2023
OpenAI’s plans according to Sam AltmanJune 1, 2023
Lawyer cites fake cases invented by ChatGPT, judge is not amusedMay 30, 2023
CLI tools for working with ChatGPT and other LLMsMay 29, 2023

Quotes

TitleDate
How to Ask ChatGPT a Technical QuestionMarch 17, 2024
MosaicML’s Open Source MPT-7B Model Writes an Epilogue to The Great GatsbyMarch 11, 2024
Gruber Responds to Gurman’s Report of AI Anxiety at AppleMarch 5, 2024
Vivek Haldar on GitHub CopilotJune 10, 2023
Back to Archive

Related Posts

View All Posts »