Michael Schade

@ajambrosino @jxnlco can we give out this whole outfit as a dev milestone kit

make me a billion dollars, make no mistakes

@TheRealAdamG @bradlightcap your gifs then and now keep us alive

@sch @bradlightcap

i'm about to cross 3 years at openai, and came across this email (copied below) i sent @bradlightcap while considering 3-month project with openai. they had the most tantalizingly insane support backlog i'd *ever* heard of after ChatGPT met the world overnight that showed no signs of slowing down, and needed a plan. i was struck by how passionately brad & @nickaturley spoke about their users describing this whole thing to me, and the weight they felt falling short of them. it really was insane, i had to triple check their accounting of it. for them solving the backlog wasn't about inbox 0, it was about trust and showing up. and although it was admittedly annoying timing—i was incredibly burnt out after shutting down my failed startup, and joining would mean telling my family they were right about my inability to abstain from computers for a year—their visceral devotion to users felt so rare that i'd booked my tickets before i even said yes. my biggest fear wasn't solving it, it's that we'd lose the forest for the trees in how we solved it. in hyperscale it's hard to realize that the real hard work hasn't even found you yet, and real hard work has a sneaky way of eradicating idealism. so i wrote the below note as a commitment to myself, and each other. reflecting for a moment on where the team is 3 years later, i can't help but feel proud of all the people who work with our customers. we've no doubt more work ahead than behind us, but in spite of this industry's tireless pace, the frontline teams have only magnified this visceral sense of purpose. some of the liveliest, most imaginative, and caring people i've ever met, learning right there with our users. 💜 ## to: openai. april 2023 ## Since we haven't gotten to know each other personally too well yet, I wanted to share a bit about how I'm seeing this and why this got my interest: OpenAI is important. We're entering an incredible time of surprise—the ways people use ChatGPT, GPT4, Whisper, and your other releases I'm sure have surprised even you all, and that's special! While we need to keep up with support because it's the expected thing to do, that's merely the baseline. OpenAI is establishing a new baseline, and should do so with the voice of your customers as well. Having a tight connection to your customer, being available to make them feel heard (and to actually hear them), and to iterate in tandem by seeing all the clever things people do with OpenAI's products (and where they hit limitations for you to solve) is key to succeeding at scale and guiding these products to the right place. I know the upfront focus is on the backlog (which is surmountable!), but I've always found that it's within these emergencies that long-term foundations and real innovation are built. Your backlog is going to be a fun challenge not just because of the current size, but because you're continuing to grow your footprint and release new products (as you should!). So in addition to getting back to the most urgent cases now and coming up with some clever ways to catch up (GPT4 should at least be a bionic arm for your team—I have some quick tool wins I'm eager to try), we'll have the opportunity to simultaneously establish our vision for that long-term support structure and pick off the highest impact pieces as part of tackling this backlog. I don't want us to get our heads above water just to be hit by another wave, so a little bit of upfront planning will go far. Put another way, this backlog is a fantastic MVP sandbox for sussing out the best long-term investments. OpenAI has an opportunity to establish support as an asset and to set a great example for the positive role AI could play in areas like customer support. I think I touched on at the end yesterday: most companies view support as drag they have to overcome just to launch new products—it's a big reason you see companies using noreply@, focusing on reducing contact rate; to survive, they bluntly stop hearing from most of their customers. By building a great support *system*, we can create an environment that's observable [0], resilient, and tunable. I hope OpenAI will be an example that automating support isn't about hearing less from people; rather, great support allows you to hear your customers better. Support should be accelerant to launching and iterating on great new products, and bringing your customers' voice to the forefront—whether it's user #1 or #1,000,000. Your product and engineering teams should be looking to your support team to hear what customers are asking for, what hacky ways they're using your APIs for clever new things, what limitations hold them back (and opportunities for OpenAI exist). I know we're in the kindergarten phase of this team, but as all parents say: they graduate high school before you know it. If I work with you on this, I want to anchor the team in what's possible when we do great work on this, as it'll be key to your success at scale and come faster than you know it. Michael [0] By the time I left Stripe, we could identify when there was an uptick in case load on an issue that made up only 0.02% of support cases but impacted important customers. And that was without GPT4…

omg clicking the supported services logos adds the CLI command to your clipboard: `stripe projects add planetscale/postgresql`

:chefs-kiss: this looks delightfully well done. although dreadfully i’ll now need to find some new busywork to pull my weight for my agents

The new OpenAI nonprofit just announced that it aims to spend $1B in its *first year" and will be led by two superb humans -- @JacobTref and @woj_zaremba. Simply put, this initiative has huge potential to do a whole lot of good. bloomberg.com/news/articles/…

@collision i’ve always thought your contributions to the cheeky pint set boded well for your future in interiors. can’t wait to visit!

Ómós, our new restaurant and guesthouse at Millbrook House in Abbeyleix, is now taking bookings for the summer. omos.co/en

has anyone given their openclaw workers an unlimited vacation policy yet?

@embirico @ajambrosino is it themes? i havent seen @thsottiaux merge this yet and i'm going to protest soon

@ajambrosino no working on something new

locking into some prompting

incredibly impressed with OpenAI Grove Cohort 2, looking forward to the next batch

OpenAI rolls out Codex Security, an AI agent that evolved from its research project Aardvark to automate vulnerability discovery, validation, and remediation (@samsabin923 / Axios) axios.com/2026/03/06/ope… #a260306p18" target="_blank" rel="nofollow noopener">techmeme.com/260306/p18#a26… 📥 Send tips! techmeme.com/contact

Codex at 2M+ active users up 25% week over week... and that was before we launched the app on Windows and GPT-5.4!

ryan and team worked extremely hard to make gpt-5.4 great for finance it's much improved for financial modeling and analysis, integrates directly into excel, and connects to factiva, daloopa, s&p global, and many more it does feel like a codex moment is coming here

open up your windows & let codex in! my fanless MINIX is about to get a lot busier in the workshop

This OpenAI blog is a must-read. Don't ask what your agents can do for you. Ask what you can do for your agent openai.com/index/harness-…

while getting a haircut today at BeauSF, stylist pulled out this AI device to analyze my scalp. it's pretty amazing. used camera + sensors to measure things like moisture, skin sensitivity, visual analysis of hair follicle quality, inflammation, etc., then recommend improvements

@guinnesschen oh my god THANK YOU

Today we shipped Handoff in the Codex app: a simpler way to move a thread between Local and Worktree.

on top of the great work in this paper, can i just say that pi has great blog design. look at the interactive encoder visual!

@bdc nice

Back to work.

broadly speaking, any unsubstantiated trend of putting ourselves first, shallow transparency efforts without follow through, etc. i pay close attention to how we distribute our tech, what happens after our internal discussions, our launch and deployment plans, etc. we are not perfect and obviously have to make tradeoffs but i think our commitment to ensuring everyone can iteratively use, learn from, and give feedback on this technology is crucial and worth fighting for. if i see that we stop making progress in this way, i’ll quit. fwiw i really do believe that society as a whole is very capable, and yet stymied by so much daily friction and asymmetric opportunity. i grew up in a very loving family but one that struggled a lot just to keep things afloat, and i’ve been so lucky to see technology change that. i believe it’s worth the hard work to distribute this technology broadly so that we can all achieve more, know more, and live a brighter future. if we lost our focus on making something better possible, we’ll have fucked up this opportunity.

@sch What observation would change your mind and make you not proud to work at openai?

i’ve known sam for 15 years, given critical feedback & seen change, and have seen him trade off personal interests to do right. while i’d love perfection, i’ve personally seen consistency in progressing our mission by actively incorporating broad input. i am proud to work here.