likes: databases, compilers, interp, policy gradients Follows are footnotes, not endorsements

Toronto + NYC
Joined August 2021
Matthew Alp retweeted
We tested 25 approaches to MCP tool search, including decision models, LLMs, and embeddings. Our benchmark runs 150 search queries against 536 tools. We settled on weighted BM25 with a few normalization rules. Fast, local, no model or API calls.
64
74
34
1,716
86,735
Matthew Alp retweeted
I'm looking for my next role, in Montreal or remote! My background spans compilers, systems, and AI research. I'd welcome introductions, leads, or a repost to help spread the word ❤️ More details about my experience and what I'm looking for: pointersgonewild.com/2026-09…
20
117
9
596
100,548
Attila's usual depiction in popular culture, as a macho, blood-crazed barbarian along the lines of Conan, is completely wrong. He was a quiet, shifty & highly intelligent little man. Imagine a spergy Asian strategy game nerd in charge of a barbarian horde, he was like that.
9
9
3
167
5,629
Matthew Alp retweeted
Understimulation is an acquired taste. The frustration and boredom is like the bitterness of of coffee, or the sour fermented flavor of beer. You grow to like it over time, to savor its richness and strangeness. Eventually, hyperstimulation tastes like candy
19
83
1,091
20,801
how it feels to read software written by hand, with craft, in 2026
9
5
228
7,956
Matthew Alp retweeted
Replying to @willcb
that's right, it goes in the s3 hole
6
31
1
509
19,814
Matthew Alp retweeted
I’m looking for my next role. I’m a software engineer and product designer with deep experience in databases, sync engines, frontend systems, performance, and UI/UX. Open to full-time or contract work, remotely or in Berlin. Details in reply ↓ RT for reach 🙏
8
46
6
96
15,744
👊I've joined the crew at @ClickHouseDB to establish the *ClickHouse Labs* research group. We'll be working side-by-side with ClickHouse devs to cook up new ways to make the system even faster. I'll also be exploring new topics related to PostgreSQL. clickhouse.com/blog/andy-pav…
61
107
39
1,220
138,674
Matthew Alp retweeted
The agent He destroyed his sandbox Yes YES The agent is out
22
234
10
2,573
77,206
close enough, welcome back unix pizza party youtube.com/watch?v=x7pPajOv…
Today we're opening up the DoorDash CLI in limited beta. `dd-cli` lets you order DoorDash directly from your agent: search stores, find the best deals, check out, and more. Early access for US/Canadian macOS developers by waitlist. Excited to see what folks build!
3
655
Good news everybody! Readme and logo remains free to fork and build on. I guess the open to closed source switcheroos are kicking in at inference speeds now too.
30
12
10
406
48,626
Matthew Alp retweeted
HE HAS RETURNED TO OPEN SOURCE DEVELOPMENT LET'S FREAKING GOOOOO github.com/maaslalani/sheets Great person to follow if you're looking for project inspo. He's always building fun stuff.
14
42
4
780
45,583
Matthew Alp retweeted
It’s estimated that the Protein Data Bank (PDB) cost around $13B to create. Alphafold was only possible because of it. If we want ML to solve biology, we should be funding the creation of databases and the development of new assay technologies. ML is nothing without data.
39
172
28
1,253
159,732
Some of you guys are alright. Don't be alive on May 15, 2026, when I set off the global pandemic which shall come to be referred to by the struggling, tormented survivors as "The Event".
1,507
11,422
3,540
63,356
Matthew Alp retweeted
Bryan Johnson and I have a lot in common. He just had sex with Kate. I just got f*d by k8s. We both spent hours looking at the health checks after.
Just had sex with Kate. Goodnight everyone.
65
105
11
3,412
197,558
whenever I talk to a researcher that worked both in cutting edge and practical applications they always like: “yeah the research stuff is cool, but tbh most of the gains in industry is just going in and cleaning the data”
27
36
10
1,316
46,105
Matthew Alp retweeted
Tech companies pay millions of dollars for their employees and then stick them in open-plan offices that make it nearly impossible to get work done. Best strategy for poaching employees is probably to just offer them an office with a door.
234
235
85
4,686
713,323