strategic GTM @SpaceXSI 🚀 prev @cursor_ai, @Alchemy
New York, NY
Joined July 2014
- Tweets204
- Following552
- Followers787
- Likes2K
Pinned Tweet
Day 1 at @spacexai.
On Friday, @SpaceX completed its acquisition of @cursor_ai for $60bn, the largest of a VC-backed startup.
Here's my thoughts 👇
It's hard to overstate the excitement shared by both teams. Both xAI and Cursor are less than four years old, and experienced rebuilds and multiple product pivots. @a16z shared that it's not often the smartest or most well-resourced team that wins, but the one that can iterate the fastest.
In <12 months our conversation with customers shifted from building a world-class IDE, to a multi-agent architecture, to first-party frontier intelligence (@grok 4.6), to agentic coworking (Grok @bot). With the addition of SpaceX's Colossus data centers, compute is no longer a bottleneck to rapidly iterating across these surface areas.
Our mission is making humanity multiplanetary. We enable that by building the best AI to shepherd us through that journey. And we get there by making AI affordable and accessible to everyone, to push the space forward. @elonmusk said to us that life can't just be about jumping from one problem to the next. You have to have something to aspire to, to dream about, to bring out your best work.
Excited to play a small part on this mission 🚀
my 2026 stack:
@bot - orchestration, acts like a digital teammate
@cursor_ai - custom demos w/o deep coding knowledge
@meetgranola - elite notetaker, alternatives aren't close
@WisprFlow - best voice transcription, barely use my keyboard anymore
@vercel - easy hosting for dynamic demos
@NotionHQ - backend for Grok Bot, what's frustrating for humans is incredibly flexible for agents
@hexanalytics, @retool, @databricks underpin my data needs
Team Bots, rather than individual Bots, will be the primary way orgs leverage Grok Bot.
Rather than having colleagues create Bots with overlapping functionality (that will vary in performance based on the user's knowledge of how to build Bots), orgs are empowering departments to build Team Bots that represent them.
Think:
Data Science > Data Bot to analyze product usage
Marketing > Campaign Bot to measure campaign ROI
Product > Feedback Bot to action on customer feedback
etc.
oh no, my wallet escaped its sandbox and ordered a $17 espresso martini. literally shaking, we need regulation
real city btw, 6 hours south of Austin.
build the future you want.
Introducing the City of Starbase Police Department's Cybertrucks!
As we build Starbase's first municipal police department under Chief Glenn Bennett, these trucks will be part of our fleet serving the community. It's been exciting to watch the department come together piece by piece, from the first hires to the equipment that will help our officers serve residents, workers, and visitors.
The department is taking shape, and we look forward to sharing more along the way.
"Building the Superintelligent Organization" was the theme of our sold out Horizon series in Dallas last night.
The main takeaway - the superintelligent org isn't the one that gives every employee an agent. It's the one that hires AI teammates to own an entire outcome.
Three CTOs shared what this looks like in production:
- @PGA: "We hover between 85 and 90% of our PRs getting merged by Divot". PMs and designers out ship engineers about 5x every week.
- @SRSDistribution: "It's the first time I've seen a piece of technology pulled from the users rather than pushed". 120 BI reports were replaced with in-context insights and a recommended action.
- @SpaceXSI likened this to being the head chef at a Michelin restaurant "You're tasting the soup, but once you know it's good, you let it go. You're not going to taste every single dish."
Thanks to @narkhip, @kevinjscott, and Cory for moderating. On to Chicago 🫡
The refutation of the Exponential Time Hypothesis is like the Warriors proving you can win championships with just perimeter shooters.
Every prior assumption of building an NBA roster around a dominant big man was proved demonstrably false
“Grok Bot can help your kid deprecate his binky”
first in my bloodline to hear that sentence 😭
Palantir cofounder @JTLonsdale says his wife's Grok Bot is helping them wean their 2-year-old off his pacifier:
"My wife has a Grok Bot at home, but she calls it the Trad bot for trad wife."
"We have 6 little kids. It's telling us about our kids' health appointments and activities, and trying to figure out my 2-year-old, how to deprecate his binky right now."
"It's telling me when to and when not to give it to him. It's pretty funny."
Via @CNBC
we’re hiring 👉 cursor.com/careers
(we’re merging websites across the SpaceXSI portfolio in the coming weeks)
If you are an ambitious young person
I still believe joining @SpaceX is one of the best career decisions you can make
The world is pivoting from bits to atoms and there is no better training ground for atoms than SpaceX
It may feel late but it's early
Also the mission! 🚀🌕
.@Tesla, in partnership with @SpaceX, has built a flying car.
The new Roadster will begin production in 2028, and the invite-only “Rocket” edition will feature a cold gas thruster system, capable of driving inverted and completing various air maneuvers.
This is a win for American capitalism. Think about how many kids that saw last night’s demo and want to become engineers now. How many entrepreneurs now believe their crazy ideas might just work. The awe and wonder of innovation is back.
What a time to be alive 🏎️
Our first X Creator Connect night in NYC (next ones are Miami and LA).
100+ of the biggest creators on X sharing how they’re using AI to edit faster, turn vision into reality in minutes, and manage high velocity posting across platforms using AI assistants.
the “jellyfish effect”
@SpaceX launched its 35th resupply mission to the ISS on Tuesday morning, with NASA capturing stunning images of the jellyfish effect caused by the spacecraft
the effect occurs when sunlight illuminates a rocket’s high altitude exhaust plume while onlookers on Earth are still in darkness
Agreed, we don’t talk enough about the shift of going from IC -> manager (of agents)
I was at an AI dinner last night in NYC.
Some interesting topics discussed
1. Performance engineers now spend most of their time managing agents and RSI loops. Most perf opts at major inference providers (together, fireworks) are found by agents.
2. RLing too much on code and math has negative effects on your model especially it’s language.
3. People were impressed by @bot and the overall acquisition of @cursor_ai by @spacexai. xAI is a legitimate frontier lab right now.
4. Most people have agents in loop writing, reviewing, and submitting PRs 24/7
5. Most agreed code review is not super productive anymore
6. Consensus was Fable 5.5 will be stunning but OpenAI will beat it in December with Bel
People will start to care less about what the underlying model is as long as you get a successful outcome
Grok @Bot will use whatever achieves the best outcome for users.
Simple questions will route to small, fast models. Questions with complex answers will route to large models.