TomerTheSecond (@tomerthesecond) header
TomerTheSecond (@tomerthesecond) profile picture

Building Voice AI Mock Interviewer. a real time Voice AI simulator for technical interviews. https://teapaw.org/t.co/YEaZN89EHE

Joined December 2013
Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)Photo by TomerTheSecond (@tomerthesecond)
Add formal verification to your code review guidelines: Here is a prompt for you: "State the property that must hold and enforce it at its owner using types, constraints, and behavior tests. Consider property based testing for broad input spaces, and formal verification or
1
19
model checking for critical logic with a precise specification. Formal verification claims must identify the checked property, assumptions, implementation revision, and reproducible checker result. Review the specification independently;
1
7
do not weaken the intended behavior to make a proof pass."
3
You can now use devx-mux inside Claude Code to pipe messages straight into your ChatGPT Desktop app using the opensource script: 👉 /devx-mux/skills/mux-director/scripts/codex-desktop-notify.ts
26
This rating is better aligned with results i actually get on real codebases than others.
3
TomerTheSecond retweeted
Download 698-page PDF >> Everything You Always Wanted To Know About Mathematics* (*But didn’t even know to ask) A Guided Journey Into the World of Abstract Mathematics and the Writing of Proofs ➡️ math.cmu.edu/~jmackey/151_12…
4
78
424
22,020
Chat is the wrong inteface for software. (Jev).
7
If you use frontier for all work, you are wasting money. Its like driving a F1 car to pick up groceries.
3
Cursor (use soley with grok) - honest, orchestrate, review, opinions, the most unbiased model i can find. Codex - Super critical things I must trust like sensitive deployment, sensitive code changes, reviews.
1
17
Claude Code - a coding workhorse not the smartest (even fable, i dont trust it).
8
You hired AI to reduce your workload and somehow became the full time supervisor of its enthusiasm.
2
I went through it all. talk to ai. specs. epics. tickets. IDE's, cli's, tmux, codex, glm, cursor, claude code, copilot, whatever. The bottom line as of today, I still have to read the code. I prefer to iterate alongside with it.
1
22
No matter how many upfront guidance etc, it will stray away. you have to hold its hands - unless its a hobby project you really dont care how its built.
1
3
Having said all it's its still ~X5 faster than doing it all by hand before AI era, and coding by hand is dead.
1