Gursharan Singh

@gdb How are you solving the code review problem for the volume of code AI is generating? Are you seeing that become a bottleneck with finite human capacity?

Software development is undergoing a renaissance in front of our eyes. If you haven't used the tools recently, you likely are underestimating what you're missing. Since December, there's been a step function improvement in what tools like Codex can do. Some great engineers at OpenAI yesterday told me that their job has fundamentally changed since December. Prior to then, they could use Codex for unit tests; now it writes essentially all the code and does a great deal of their operations and debugging. Not everyone has yet made that leap, but it's usually because of factors besides the capability of the model. Every company faces the same opportunity now, and navigating it well — just like with cloud computing or the Internet — requires careful thought. This post shares how OpenAI is currently approaching retooling our teams towards agentic software development. We're still learning and iterating, but here's how we're thinking about it right now: As a first step, by March 31st, we're aiming that: (1) For any technical task, the tool of first resort for humans is interacting with an agent rather than using an editor or terminal. (2) The default way humans utilize agents is explicitly evaluated as safe, but also productive enough that most workflows do not need additional permissions. In order to get there, here's what we recommended to the team a few weeks ago: 1. Take the time to try out the tools. The tools do sell themselves — many people have had amazing experiences with 5.2 in Codex, after having churned from codex web a few months ago. But many people are also so busy they haven't had a chance to try Codex yet or got stuck thinking "is there any way it could do X" rather than just trying. - Designate an "agents captain" for your team — the primary person responsible for thinking about how agents can be brought into the teams' workflow. - Share experiences or questions in a few designated internal channels - Take a day for a company-wide Codex hackathon 2. Create skills and AGENTS[.md]. - Create and maintain an AGENTS[.md] for any project you work on; update the AGENTS[.md] whenever the agent does something wrong or struggles with a task. - Write skills for anything that you get Codex to do, and commit it to the skills directory in a shared repository 3. Inventory and make accessible any internal tools. - Maintain a list of tools that your team relies on, and make sure someone takes point on making it agent-accessible (such as via a CLI or MCP server). 4. Structure codebases to be agent-first. With the models changing so fast, this is still somewhat untrodden ground, and will require some exploration. - Write tests which are quick to run, and create high-quality interfaces between components. 5. Say no to slop. Managing AI generated code at scale is an emerging problem, and will require new processes and conventions to keep code quality high - Ensure that some human is accountable for any code that gets merged. As a code reviewer, maintain at least the same bar as you would for human-written code, and make sure the author understands what they're submitting. 6. Work on basic infra. There's a lot of room for everyone to build basic infrastructure, which can be guided by internal user feedback. The core tools are getting a lot better and more usable, but there's a lot of infrastructure that currently go around the tools, such as observability, tracking not just the committed code but the agent trajectories that led to them, and central management of the tools that agents are able to use. Overall, adopting tools like Codex is not just a technical but also a deep cultural change, with a lot of downstream implications to figure out. We encourage every manager to drive this with their team, and to think through other action items — for example, per item 5 above, what else can prevent a lot of "functionally-correct but poorly-maintainable code" from creeping into codebases.

@AmazonAlexa whisper mode is 🫶. Small things done right for users

Hongly's healthcare was fully funded on @watsi! Help support many such children with beautiful promising lives ahead, in need of care. watsi.org/profile/b183c9…

Unbelievable how hard it is still to organize your apps on iOS. The “app library” organization is junk, and the Tetris you’ve to play to manually organize folders hurts the thumb. Where are the product masterminds at @Apple

@RyanHoliday That is one option I'm considering after reading it religiously this year...unless you've another similar recommendation Ryan from the stoic bookshelf

Anyone planning on reading (or re-reading) The Daily Stoic this year?

PSA: if you use MacOS’ Calendar to manage your time/schedule, you’ll be left behind in life.

And it continues…service brought a replacement post that had way more cosmetic issues. Apologizes but calls it within “spec”. How on earth is this company operating? A car with a battery and software is…no…longer…a normal car?

Brand new @Tesla delivered midnight 6/30 to meet sales quota-5 cosmetic 1 structural 1 exterior defect. Refused car replacement, service-“it’s normal we’re rolling ‘em out like peanuts”. Mobile svc comes home, has wrong parts. Frustrating! @elonmusk you still running the co?

Aurelius on being present and living in gratitude — “no one ever loses a life other than the one they are living, and no one ever lives a life other than the one they are losing”. From @RyanHoliday’s The daily stoic

@joulee From the one characteristic panning the universe - change.

What is the source of human ambition? Does it from of insecurity? A need to prove something? From love?

@AJamesMcCarthy How large is it approximately?

This is how much the comet moved in just 10 minutes tonight. This thing is FLYING. Must be moving close to 100,000mph right now.

@urbanturbanguy @jvalaaa @cdsinghneo @_karandeep @kulmeets This group, any day, this topic, mmmm

@jvalaaa @gursharan @cdsinghneo @_karandeep @kulmeets @gursharan can you host us for dinner Friday ? :)

playing around / thinking about chat gpt 🧐

@urbanturbanguy @jvalaaa @cdsinghneo @_karandeep I haven’t seen A LLM model yet trained on another language. The corpus and language modeling is the challenge, not impossible

@jvalaaa @gursharan @cdsinghneo @_karandeep if we were to fund a couple of students to bootstrap GPT work in gurmukhi. What would it take? Is that even a technical possibility as of today?

@AJamesMcCarthy Surreal! Space is so recalibrating to the mind engrossed in the everyday triviality. Thanks for sharing as always!

"What does it look like before edits" is a common question. It looks like this. This is a raw unprocessed image. It's like undeveloped film. Processing is done manually because I can't count on automatic processes like a regular camera. The entire image is inside this 32bit TIF.

The light from these galaxies colliding was emitted 23 MILLION years ago, and then travelled across intergalactic space only to be caught by my telescope 3 feet before hitting the ground. I get an existential crisis when I think too much about the mechanics of my hobby.

Happy New Year! While many of us celebrated by texting our loved ones on WA, there are millions of people in Iran and elsewhere who continue to be denied the right to communicate freely and privately. So today we’re making it easier for anyone to connect to WA using a proxy.

@stevecoots @MikeDavisEsq @la_vfx @BCineMag @netflix Wow on the work that went behind the closeups of Stella, Rocky, Poloma and all other wonders. My fav visual amongst many others was the creation of H from the tentacles of energy swirling inwards. Magnificent! Thank you.

An interview I did with @MikeDavisEsq and @la_vfx for @BCineMag about #OurUniverse Streaming now on @netflix britishcinematographer.co.uk/how-lux-aetern…

@Theholisticpsyc May you achieve all you want. Keep sharing your wisdom.

2023 energy: forgiving myself over and over again. Allowing myself to be human. Loving myself through mistakes. Being the person my younger self needed.

Rarely does one come across content that can open new pores of the brain. Our universe on Netflix in sir @morgan_freeman’s voice was one such for me. netflix.com/in/title/81243…

@SahilBloom And zooming literally out and see self as a speck on the speck that is the blue dot in a vast universe, helps recalibrate in a different way

Being perpetually zoomed in creates two challenges: 1. Struggle feels bigger than it really is. 2. Growth feels smaller than it really is. So zoom out! The 10,000 foot view provides perspective—on the manageable nature of your struggles + the impressive nature of your growth.

We live our lives zoomed in. Zoom out regularly to reclaim your perspective.

@pitdesi when you wear a turban, that q in America often ends up verifying if your *really* are Indian. That stings, for the love of all humanity. I do not mind at all, though, and often feel like wrongly answering them to see their reaction

Some people don’t like it when people ask where they are *really* from and I don’t understand I grew up in Pittsburgh and am of Indian Rajasthani heritage, and am happy to talk about any of those. Can someone who takes offense to the q help me understand why it’s offensive?