Happy International Women’s Month!
Life update:
As many of you who have wondered where I’ve been the last few months:
I've spent the last few months heads-down building a ton of new stuff that I’m releasing soon (websites, apps etc.).
1. Releasing New Website - // erinjerri.com is currently being updated in spurts. 🌐
Outside of some basic content of past experience in my portfolio, the official site will be dropped soon. It is * ALMOST * done, I am finalizing my store and three.js animations and an internal dashboard currently with PayloadCMS. I’m using AI to try to create the dream workflow, making sure this template will work so that it can easily be deployed across 4 websites (thanks to a lot of help by upgrades with OpenAI Codex). This is how you use AI effectively (templatizing your work and having it extract the base fonts and colors - mood board essentially) so you only need to spend time uploading content (AI is so magical like that). I’m using this website as the base for app marketing website, the non-profit, and a fitness website (working on a friend’s side hustle body building business).
2. AI Benchmark/Evaluation for a Computer Use Agent - Spatial Computing and AI 👓
3. New AI x productivity app
This benchmark is a branch paired with a new beta app I’ve been stealthily working on for a productivity x AI app that works in spatial computing/AR VR/AI glasses. Release coming soon! If you’re signed up to this Substack as as subscriber you essentially are on the beta tester list!
The Mac ecosystem is coming first with iOS and VisionOS, with other platforms to follow. Everyone wants this who I’ve talked to in the last two years, demands I have this on Android, Windows, and Linux (it is A LOT of work to make this cross-platform and I’m starting with Mac, Quest, and AI glasses, be patient people!).
In this issue:
A. My AI Tech Stack
The gist: For both software engineers and non-technical people. I get asked all the time by non-technical friends what I use besides ChatGPT and while there’s an explosion of apps and many new ones released everyday, here are some of my tried and true apps I like to use daily or weekly at the very least (in combination with other apps like Figma for design and programs like that). Scroll down for more details.
B. 3 Mistakes to Avoid When Coding with AI - for software engineers
The gist: I see a lot of people producing AI slop (spaghetti code) that doesn’t work in production and junior engineers who have no idea what it is they’re writing in command line or writing in their Integrated Development Environment (IDE), their code editor. While we give a lot of power to AI agents and automate a lot of mundane work, there’s some core engineering that is being missed. Don’t overly rely on AI agents, you still have to learn how to do long division in school (and don’t always have a calculator on you). Scroll down for the more details.
C/ My Recommended Books on AI + Data Science Industry Leading Computer Scientists, Software Engineers authored by women of color - for engineers and non-engineers
Future issues will include:
- My software engineering and design process (building in public) across VisionOS, iOS, watchOS and more,
- How to batch with AI has a productivity hack for ADHD, and
- A long-awaited post on longevityAI as it intersects with women’s health.
Extra Bonus:
For all your productivity junkies like me, I share a brief post on productivity (inspired by Taro (YC Backed company) Co-Founder and formerly Meta Engineer, Alex Chiou and courses for those getting hired in Machine Learning and AI as software engineers.
Enjoy!
Sincerely,
Erin
A/ My 2026 AI App Stack 🤖
Since folks have asked me for ages to have a more technical AI post (but doesn’t completely go over your head), here’s the long-awaited post on my current AI tech stack across consumer, software engineer/developer, creative, and design tools.
I use different tools for different jobs: research, coding, product planning, notetaking, and generative UX exploration.
This is the setup I’m using daily right now (and some a few times a week for design), including what I trust most, what I test often. Also, for many who have also asked because hardware is all the rage, I’m still figuring out how to use OpenClaw * with caution . *
The Standard Foundation Models
- OpenAI ChatGPT - // the OG and still most performant
- Anthropic - Claude - // the most ‘safe,’ was the only FM to hallucinate, I wrote a book I never wrote.
- Perplexity - // most up to date with real time data supposedly
- Google Gemini - // best for search and references
- Microsoft - CoPilot - // mostly hyped because of OpenAI, decent
- xAI - Grok - // surprisingly good and very specific
- Pi - // the most personal
- Poe - // another one that gives short and snappy responses
Best for Programming/Coding
- Cursor - the game changer that moved everyone from VSCode and Copilot.
- Also XCode using GPT
Best for Product Management
Other tools I’ve used:
Want to try:
Best for Notetaking Meetings and General Work
Best Generative AI / UX Tools I Actually Use: How I test 10+ diffusion models quickly
- Ideogram and Reve - These two are definitely prettier than me using ChatGPT, Sora, Nano Banana. Easier to use if you don’t want to use MidJourney and Discord or other apps (back in the day we had to write StyleGAN from scratch!).
- Aura.build - my favorite app by Meng To, who I have long been a fan of (he is super nice over email btw). His site designcode.io has evolved a lot over the years (which came into being after I bought his first book which was about learning mobile development and a front-end engineer and designer and is now a go-to for designers learning anything technical). I used it to create an entire site (without putting it into full functional production, tried it in Payload it still cant quite completely integrate into the back-end with Cursor, but it’s a great prototyping tool to see a single page websit.
B/ 3 Mistakes to Avoid When Coding with AI - for software engineers 💻
- Becoming overly reliant on the agent.
When you surrendering all control to the foundation model and let the the agent run rogue, it can easily make mistakes and hallucinate.
While we have AI agents, humans still have agency and need to be able to do more than guide, but steer and direct what you want to make happen.
- Being too delusional thinking ‘vibe coding’ will * magically * create your app for you.
// Read the docs (esp when doing open source - OSS).
Plan and don’t do it all by yourself.
I use a combination of Notion, a foundation model of choice (I like to use ChatGPT and Grok for this) where I can product manage in Notion, have ChatGPT and Codex give me honest metrics about how feasible it is to complete.
- Underutilize AI for writing documentation.
// Most software engineers I know have really poor communication skills, don’t like writing, either over comment or under-document their process, making it harder to hand off code, conduct code review for quality, or debug (find where a bug is).
C / My Recommended Reading List - International Women’s Month Edition:
Top Books on AI By Software Engineersa and Computer Scientists - Women of Color Authors 📚
As a fellow technical author and software engineer, I share my shortlist of top industry authors who are all leading computer scientists in AI I’ve had the privilege of meeting. I am greatly inspired by their work and what they’ll do next.
Fei-Fei Li - author of The Worlds I See
This is a good inspirational read and friendly to read for non-technical people. Known widely as the Godmother of AI and creator of ImageNet, Fei-Fei Li is short of no accolades as also a Stanford Professor in Computer Science, founder of startup WorldLabs working on spatial intelligence and was past Chief Scientist of Google as well as the founder of Stanford Human Centered AI (HAI). It was such an honor to meet someone who has long inspired my journey in AI since auditing CS 231N online (when her PhD student at the time, Andrej Karpathy) was the graduate student instructor for the class almost 10 years ago now.
Me with Professor Fei-Fei Li at Stanford Book Launch (at the Cybersecurity Policy Center - CPC)
Chip Huyen, author of two O’Reilly Media books, AI Engineering: Building Applications with Foundation Models (published in 2024) and Designing Machine Learning Systems; An Iterative Process for Production-Ready Applications (published in 2022).
This is a must-read for software engineers and data scientists, and it’s actually ranked as the most popular book by my publisher, O’Reilly Media, on their online platform, Safari. She is a Vietnamese American author, Stanford alumnus, past startup founder, and a writer.
It was so nice to finally meet her at PyTorch Conference last year and seeing her survey how people are using LLMs today.
I’ve long followed her great blog, which you should read
here.
As a fellow O’Reilly Media author, I love plugging other O’Reilly Media authors. Check them out here.
And last, but not least, you should read
Unmasking AI, by Joy Buolamwini
Widely known for her work, “
Gender Shades” as as the founder of the Algorithmic Justice League, Joy has tackled the issue of bias highlighting disparities in intersectional identities (finding that dark skinned women of color were not accurately tracked as well as cis straight white men).
BONUS: Taro Course on Mastering the Machine Learning Interview 🍠
I took this course on Taro last year, and it had some great material presented by my friend Yayun Jin. She runs through practical examples, some other acronyms that help you structure your interview responses, and closes the gap in the industry, providing really useful information that has been daunting for many software engineers and data scientist for years.
Ξrin’s Jerri’s Newsletter is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Subscribe
Follow
@erinjerri for more on AI, XR, web3, productivity, and technical/design process.
