Jeremy Kun
banner
jeremykun.com
Jeremy Kun
@jeremykun.com
Portland-based mathematician and software engineer. Building a homomorphic encryption compiler at Google.

https://jeremykun.com
https://pimbook.org
https://pmfpbook.org
https://buttondown.email/j2kun
https://heir.dev
I found out about this about a year before sending my oldest kid to kindergarten and I started doing phonics with him right then and there. Turns out our school is fine, though.
November 11, 2025 at 6:27 PM
Indeed, this seems ripe for automation.
November 10, 2025 at 2:54 PM
If they made a feature length film where Link has NO lines beyond grunts, that would be...incredible.
November 7, 2025 at 7:07 PM
But I feel like if you gave an unsorted set a "pop" method, that would be the FAFO queue.
November 6, 2025 at 8:35 PM
This has big "send tweet" vibes jeremy you can do better
November 6, 2025 at 8:33 PM
I like the lowercase i here.
November 6, 2025 at 4:22 PM
I can edit the original post on mastodon, but not the screenshot🙈
October 30, 2025 at 4:26 PM
Yeah, clearly it needs to be induced
October 30, 2025 at 4:26 PM
I think it's just because they like them and they're happy to see them.
October 27, 2025 at 1:14 AM
I honestly want to hear your answer, but in my experience the people who treat LLMs like they can write and understand code well are people who don't do a lot of software engineering. Using an LLM effectively for coding requires an INCREDIBLE amount of guardrails, guidance, and human in the loop.
October 21, 2025 at 3:55 PM
Do you code regularly with LLMs?
October 21, 2025 at 3:53 PM
I still have a lot of gaps in how these algorithms are used for useful things in typical compilers, and the stuff I'm using it for is rather basic, but I do want to start going through the algorithms, in part to finish upstreaming them to MLIR. Might be a while 😅
October 20, 2025 at 1:04 AM