Mike
185 posts
@__Tkat__ @internalises arguably worse might as well switch to olipop vintage cola atp
English
Pump fun was always bot wars, what is there to sue for? The game is the game
Burwick Law@BurwickLaw
Pumpdotfun & Solana lawsuit update: Leave to amend (file new complaint) GRANTED “What appeared to be a fair, automated marketplace was, Plaintiffs say, structurally tilted to extract value from ordinary users while rewarding those with privileged access to Solana's infrastructure and Jito Lab's transaction ordering tools.”
English
Want to join the team that works on this?
We’re hiring a Platform Services Engineer at Browserbase. Sneaker bot / anti-bot experience is a major plus.
jobs.ashbyhq.com/browserbase/14…
harsehaj ⋆˙⟡@harsehaj
We forked Chromium and broke it until it worked exactly the way we wanted it to. Chromium wasn’t designed for long-running server workloads, CDP consistency under automation, or removing identity leaks. So to build real browser infra for AI, we had to reshape it from the inside out. Read our deep dive on how we patched Chromium and built a custom binary.
English
Mike 리트윗함
Yapping in the video below but if you're hiring engineers or interested in engineering roles / who's hiring here's the tl;dr:
This software solves
1/
Problem for employers:
Too few applicants = not enough to hire
Too many = no time to review
Either way, great candidates get missed. Inbound becomes noise.
Via semantic search + boolean screening at scale you can surface candidates that are a good fit at scale. Results are 10x better than alternatives that are just keyword matching, we consider context.
2/
Problem for candidates:
Even if you're great, your application is one of thousands. Recruiters skim resumes in seconds. No signal = no response.
Because we make search + surfacing best talent extremely easy for employers you get better access to the roles that make sense.
3/
What we’re building:
Structured flows for every inbound applicant
Signal detection across resumes, portfolios, social links, GitHub, etc
Auto-prioritization based on quality, not keywords
4/
Why it works for hiring teams:
It turns unstructured inbound into ranked, high-signal leads. Founders or recruiters only review the best.
More signal, less noise, no wasted time = more interviews.
5/
Why it works for candidates:
You're not lost in a pile
Your work gets parsed, scored, and shown based on quality
Distributed to high quality teams
Using it now to hire engineers in NYC and SF. Will expand to more roles + teams soon.
DM if you want early access for hiring. If you're interested in who's hiring check out the job postings below:
English
Mike 리트윗함
@i7solar @GeckoSquadron @LicketySpliket and I worked on a pump.fun bot a bit back, could be useful mikem.codes/if-you-aint-fi…
English
Mike 리트윗함
we hit it.
just reset the counter to $10B. back to work.
Guillermo@Delmo_dev
We're about to hit $1B processed on Whop. I added a live counter here if you want to follow: whop.com/leaderboard
English
Mike 리트윗함
With all the AI-powered job interview tools out there, the way we hire software engineers needs to change.
Here’s how I would do it:
Step 1: No More Trivia Questions
I’d skip trivia questions entirely. They’ve always been pointless, and with AI, they’re even more irrelevant.
Instead, I’d start by asking candidates about a project they’ve worked on and really dig into the details. If someone is making things up, it’ll be obvious when they struggle to answer specifics like:
- “How exactly did you implement authentication?”
- “Can you give me examples of SQL queries you had to optimize? How did you improve them?”
A real engineer will have no problem explaining their past work in depth.
Step 2: An AI-First Real-World Coding Challenge
Next, I’d give them a larger project. Something like building a simplified version of an existing SaaS product and seeing how far they get.
They’d have one day, complete freedom to use any stack, and full access to AI tools.
A good full-stack engineer using AI should be able to get something basic up and running in a day. At minimum, a simple auth system and core functionality.
Step 3: Code Review & Understanding
Afterward, we’d go through the code together. The goal isn’t to check if they wrote everything by hand (AI-assisted coding is expected), but to see if they actually understand what’s happening. They should be able to explain the logic well enough that you’d trust them to debug and maintain the system if something broke.
————
This process is way more involved than the usual “LeetCode-style” interviews, but it’s both AI-proof and AI-first. More importantly, it gives you a much better idea of how someone will perform in a real-world engineering role.
English