
cs 29’ @umich | SDE Intern @AWS | @V1Michigan
Ann Arbor, MI
Joined October 2025
- Tweets281
- Following482
- Followers137
- Likes3.7K
I've always wondered why AI labs don't just do this. Save all the compute you have to train better models and own all the output of your models for yourself
I getting a feeling that OpenAI and Anthropic soon will ditch customers completely, just use their models 100% internally and go all in on frontier science and engineering research
Like eg if they actually solve quantum gravity or cure cancer, and own the right to license the solution, can you imagine the ROI behind that? No more listening to users bitching about usage or model nerf or whatever
Please opensource 🙏
We've decided to sunset Arc on December 22, 2026.
That's not an easy sentence to write. We spent years building Arc and celebrating each and every release along the way: we have so much pride and love for what we made alongside the Arc community.
Our priority is to deliver best-in-class security for our members. Keeping a browser secure takes constant, dedicated work that is only increasing in the age of AI. Dia was built from the ground up to meet today’s security standards, so that's where our engineering and security teams will focus moving forward.
We've worked to make switching to @diabrowser as easy as possible for Arc members. Here's how:
· Your passwords, bookmarks, extensions and history come over to Dia seamlessly.
· Spaces, Folders, and Tab Groups are in Dia and import over with your data intact.
· Dia now surpasses Arc benchmarks in stability, speed, and security.
· The browser features you use in Arc are free in Dia, and the AI is optional.
Thank you for using Arc, for pushing us to make it better, and for building alongside us. The team that made Arc is the same team behind Dia, and some of our most ambitious work is just ahead. We'd love for you to be there with us when it lands.
Dylan Li 〽️ retweeted
I reverse engineered Pocket and made it 10X better.
Right now: the app is full of ads, the microphone is low quality, it constantly misspells peoples names, and has no agentic features.
I changed that.
Waterga(i)te
We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work.
At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated.
We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.
My @ChatGPT dot just sent an email without my knowledge idk what to think of this
This is so smart
if you’re in b2b
PAY ATTENTION what Ramp is doing with their outbound messages.
• email 1 congratulates the prospect on a new funding round and asks for 15 minutes.
• email 2 opens with "Shoot! Forgot to put this in my last email."
i'd bet money that “slip” is written into the sequence on PURPOSE.
the prospect gets a 2nd touch with a reason behind it, from someone who seems human.
absolutely class.
Last weekend at @mhacks, @manavkhosla25, @varchaskukreja, and I built Physical Realities, a Jev-powered simulation platform that lets operators test how changes to physical venues affect customer experience. We ended up winning first overall! Check it out below:
Dylan Li 〽️ retweeted
My team and I built Mack at this past weekends @mhacks (which won 🥇 place and $2500 for the AI track)
Mack is a voice AI Chrome extension that helps the elderly and those with accessibility issues simplify and navigate confusing websites.
Check it out: devpost.com/software/lens-ai…
🚀
excited to host @spacexai with @v1michigan today!
if you're at @umich, come stop by
luma.com/0ku8agev?tk=x3WVj3
Dylan Li 〽️ retweeted
Good point from @borrasque_ here. The truth is that orgs like @V1Michigan make mistakes.
We hope our hit rate is high, but we know we’ll miss ambitious people who go on to do great things with or without us.
To this effect, we’re trying to open source as many resources as possible and invest more time in community events (still a work in progress).
A rejection doesn’t have to be the end of someone’s relationship with V1. We’ve had people we’ve passed on come back, prove us wrong, and even join our team to help scale the org. See my V1 rejection from freshman year attached :)
To anyone we’ve missed who’s still doing great things on campus: we’d love to have you around more.
And to you, I’m excited to see how your gap goes (I know I loved mine). Glad you’re proving us wrong.
during my freshman year at umich there was this prestigious startup club i really wanted to join
all my friends were talking about it like it was the place to be
i remember asking some of the members what type of person they were looking for, and a few of them told me "someone who has agency"
i thought to myself "that's lowk me"
i got rejected 3 rounds in during my first semester. it was a coffee chat/interview round and i guess i didn't do as well as i thought
i applied again my second semester. rejected in the first round. it felt like everyone in that club was rooting against me
but here i am, founding engineer at a yc-backed startup during my sophomore year
what surprised me the most is almost everyone at that startup club recruited for corporate
just because some people don't see potential in you doesn't mean you don't have any
maybe you're just not in the right place, or surrounding yourself with the right people
Prediction: AI PE firms are the future. Traditional SAAS companies will become easily replaceable as the cost to produce software goes down. However, the moat of an AI PE firm is in how well they can utilize technology to multiply the output of traditional companies.
Plus, AI PE firms hedge against the risk of SAAS value collapse because they own the companies beneath the technology. Additional valuation is created by the software around the company, but the baseline is still high because there's ownership of a genuine company underneath.


















