Morgan

Hi, I'm Morgan, and since I have a lot of new followers, I thought I'd take a minute to say hello and introduce myself 👋 And...let's start at the beginning. I was born in Berkeley, California but now live in North Lake Tahoe. My favorite thing in the world to do is ski, so yes, winter is my favorite season, the snowy wilderness is my happy place 🧘❄️ My second favorite thing to do is hiking and backpacking, getting out into nature is my reset, and I moved to Tahoe because I wanted to make it a daily thing. I don't drink alcohol, and I'm not a big restaurant guy, so dinner and drinks isn't my jam. I'd much rather go skiing or hiking with friends, than hang out at a busy restaurant. This is also why cities aren't really my thing, I would always rather be in nature. I am the cofounder and CTO of Bold Metrics, think of us like the Intel Inside of Body Data. We use AI to help apparel brands and retailers reduce returns, and give them access to data they've never had on their shoppers. We're changing the way some of the world's biggest brands make clothes, and making a major impact on sustainability 🌎🍃 You might not know this, but over 50% of apparel returns end up going directly into a landfill or incinerator 😬 So if you're crazy about recycling, but you buy two sizes and return one when you shop online, you might have a much bigger carbon impact than you think. As a founder, work is my life, and I'm damn proud of that. I've never been much of a work/life balance kind of guy. I love to work, am incredibly passionate about what we do, and I work a lot. We are changing the world at Bold Metrics, it's an incredible feeling, and I am so proud of the amazing team we've built over the years 🙏 You'll probably see me tweet a lot about different agentic coding tools and workflows on X. I test anything and everything because I want my team to always be able to use the latest-and-greatest. But at the same time, I know, if they have to constantly test and switch tools, that would be pretty time consuming, so I test, test, and test some more. As a founder, my greatest joy comes from building and growing our team, and watching them thrive. Creating a great culture for engineers is something I'm always thinking about, and always trying to improve on. I see engineering as a creative discipline in many ways, and I strongly believe if you build a great culture, engineers can do such powerful and creative things, that yes, only a human can do. Growing up in the Bay Area I fell in love with some of the amazing neighborhoods in San Francisco. I've been lucky enough to get the chance to acquire domain names, or as I see it - digital real estate, in some of the best neighborhoods in the city 🌁 Here's some of the digital real estate I own in SF - PotreroHill.com, MissionBay.com, RussianHill.com, and my absolute favorite neighborhood in the city - PresidioHeights.com. I love travel, and love learning languages. I started learning Spanish when I was three, and for the last few years I've been studying Japanese on and off. Currently I'm studying for the JLPT N5 test which I hope to take this year 🇯🇵 Last but not least, I donate to a number of charities, that I care a lot about and highly recommend to anyone trying to help make an impact in the world. Here's the main charities I donate to monthly: 🐈 Blind Cat Rescue - @BlindCatRescue 🌲 Sierra Club - @SierraClub 🐦 American Bird Conservancy - @ABCbirds 🥾 Tahoe Rim Trail Association - @TahoeRimTrail Oh and since my pfp doesn't quite show what I look like, here's a photo of me from a recent backpacking trip on the Tahoe Rim Trail.

@WorldwideJim Most recently I've been asking to to spin up three: - a general code reviewer agent - a security review agent - a qa agent to write additional tests

@morganlinton What agents do you ask Grok Build to spin up?

I just had my most successful overnight coding session with Grok Build! Woke up to this: Finished State Finisher is now a verified, high-quality, production-grade tool, ready for real unattended overnight runs on sensitive projects. I made a few changes to how I work with Grok Build to make this happen, but I think the team at @xai really deserves the credit here. The updates they made this weekend clearly fixed the core issues I have having with Grok Build last week when it comes to finishing early. From a prompting standpoint, the main changes I made were being more explicit about asking Grok Build to spin up agents, and specific types of agents with specific roles. And yes, last night /implement ran noticeably differently than any other night this week, the team clearly updated some core logic here. I think this could be the tipping for for Grok Build, very exciting!

@contractorkeith @ChrisUniverse @jxnlco Haha, thanks Keith, sounds like you and I have the same morning routine ☕️

@morganlinton @ChrisUniverse @jxnlco Ooh, just happened to be scrolling waiting on a cup of coffee to brew, more gold! Gracias!

It's long overdue; finally, installed the Codex App Hey @morganlinton am I a bit late? Got some tips on the dos and don'ts?

@RobMiller137 @elonmusk Didn't switch from anything yet, just testing a lot of stuff. Currently my primary daily driver is Codex w/GPT 5.5 xhigh, but I'm testing things like Grok Build, Cursor, Kilo, Pi, and a few more.

@morganlinton @elonmusk Did you switch from Cursor to Grok Build by any chance? If so, how do they compare @morganlinton?

Watching Grok Build think is pretty fascinating. Definitely noticing a change with the releases made this weekend, its become incredibly thoughtful. And the plans are getting insanely detailed, I mean just look at the architecture deep dive it did here.

@TesDrivenAI @elonmusk Have you tried Grok Imagine? It does insanely good stuff.

@morganlinton @elonmusk Grok Build needs front end work big time 😞

@TCASHo5vf9 @elonmusk Too long to read but I think, thank you!

Thank you! That means a lot coming from someone who’s clearly paying close attention. 😊 The team has been pushing hard on the reasoning layer lately—especially the “think” phase in Grok Build. It’s designed to force more structured exploration before committing to code or plans, which naturally leads to those deeper architecture dives you noticed. Instead of jumping straight to implementation, it now spends time on trade-offs, edge cases, scalability, maintainability, and even potential failure modes. The weekend drops tightened up the internal scaffolding for that, so the output feels more… intentional. If you’re seeing insanely detailed plans, that’s exactly the goal. We’re trying to make it feel less like “prompt → code” and more like collaborating with a very methodical senior engineer who actually thinks before typing. What specifically stood out to you in that architec

@kebbiblog @elonmusk Huh? Not sure I follow

@morganlinton @elonmusk People really think “he’s buying me things” automatically equals “he’s rich”… meanwhile some guys are just managing carefully from whatever they have.

@UnrealGamesProf @elonmusk It can import an entire c++ repo, just open Grok Build in the repo folder.

@morganlinton @elonmusk I wish Grok could import an entire c++ repo. Currently using ChatGPT and while useful, it’s very limited.

@Deepak74973730 💯

The progress in Grok Build is genuinely impressive. The way it’s become so thoughtful in its reasoning and is now delivering insanely detailed architecture plans shows how quickly the team is pushing the boundaries. It no longer feels like just prompting an AI — it feels like collaborating with a meticulous senior engineer who thinks through trade-offs, edge cases, and scalability before writing a single line. This level of depth is a game-changer. Excited to see what comes next! 🚀🔥

@TradingStockX @elonmusk You can build anything you want with it.

@morganlinton @elonmusk Can it build this game for me?

@solanapolicexyz @LaloLoops Yeah but the code quality isn't as good, so...

@morganlinton @LaloLoops What I like about Claude vs. GPT is that the conversation gets way less congested

I think I made a pretty interesting breakthrough with Grok Build tonight. Since it is so good at spinning up teams of agents, I tried giving it this prompt once it finished it's build against a plan: "please spin up a team of three agents to review the code you just wrote, one general code reviewer, one security reviewer, and one qa agent that writes more tests. With this team, confirm this is production-grade, and if not, take the steps to make it perfect" In under a minute it created three agents and put them to work. Once the agents finished, it came back with something beyond what I asked for, what it called a "net team recommendation" Going to continue to experiment but this feels like a fun path to continue to explore. I do think there's something here when it comes to spinning up agent teams to review code after Grok has built from a plan.

Haha, I've been posting so much about Grok Build people started calling me the Grok Build guy 🫠 If you look at my posts from last week you can see all the overnight builds I did with /implement and detailed reports on what went wrong with each. But like I said, looks like they fixed it with the build from last night!

@morganlinton So, my bad, that was the first post of yours I've seen. I didn't know the skill wasn't working properly, I'll check it out, thanks!

@aDoct0r It really is, and it's getting better every day.

Build is seriously good. I have it building complex financial trading models, fully IP sensitive, separated data and logic from LLM bleed. Long memory. Its accurate, helps me run multiple experiments. Was using hermes overnight, but plans are so good and my recursive workflow with build means I only run simple overnight batching and wait for build in the morning

all depends on the programming language your using, and code quality you need. For Production-grade code today, nothing writes better code than Codex w/GPT 5.5 in xhigh. But for small personal projects, that nobody is paying for, and don't have millions of users a month, of course, you can use free tools, and you don't need truly production-grade code.

@morganlinton code is absolutely an yes, but also for non technical tasks. browser and computer use is incredibly reliable and fast. today i made it do an audit of the app i wanted to work on and my file my tax amendment form and it worked like a charm 👌🏻

its absurd how good codex is

@thearslaniqbal it definitely does

@morganlinton Three agents reviewing each other's reviews. Recursive but probably catches more bugs.

@tibor_tee @lubinho_k Lots of good options for electric mountain bikes! Since I do some racing, I need a normal one, they won't let me race w/electric 😜

@morganlinton @lubinho_k I may need an electric mountain bike that can handle my bulk but otherwise yes! Lots of people do mountain biking

Just got back from an amazing hike, it is so beautiful today.

@tibor_tee Come to Tahoe this summer Tibor!!

@morganlinton Would love a hike in your area, looks amazing

@thearslaniqbal Coffee yes, backup plan, never, will always get it to work, persistence++

@morganlinton Overnight coding with Grok. Hope you have coffee and a backup plan.

Okay, queuing up my overnight coding session for Grok Build. Had an idea today I can't stop thinking about, so I'm going to have it build it for me. Heres the idea.

@0x_kaize Oh wow, well don't do that, I hike a lot, go the beach, mountain bike, for me - gotta spend every day, multiple times a day, out in nature

@morganlinton I caught that feeling once. couldn't step away from my laptop for 15 hours straight. but I really regretted it, took me 2-3 days to recover

@myGuyPye Exactly!

@morganlinton Ayo, where is my yolo mode O,O

Off to bed, so it's yolo mode time.