former FW & SW Eng, now I&C Eng. #!/bin/sh ‹self purge after follow›

muterin Jateng
Joined August 2018
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
I wrote a JVM build system in Rust. ☕🦀 jrs brings a Cargo-like workflow to Java: one jrs.toml, lockfile, single binary, no scripts. Java, Kotlin, Scala & Groovy. Maven Central. Fat JARs, jlink & GraalVM native. Experimental & OSS. getjrs.dev #Java #Rust #JVM
🤖 Made with AI
1
1
8
268
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Semantic Diffs are coming to @rescriptlang 🧵 Review a compiler change as the change it makes to the compiler's own objects: here, the parse tree of `%raw`, Before and After.
2
7
2
44
2,637
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Di sinilah local model berguna, pakai Qwen-3.8-27B ablit gak hanya bisa disuruh nyerang, kadang malah ngasih saran pendekatan yg paling kejam dan brutal
Mau hardening program: ✅ Mesin buatan sendiri ✅ Program buatan sendiri Lalu nyuruh clanker nulis kode buat attack mesin dan program sendiri buat verifikasi keamanan 💀☠️ Kena safety check UDAH DIBILANG PUNYA SENDIRI SEMUA 😌💀☠️ Tetep kena safety check LOL
2
5
1
178
8,380
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Welcome to the world @OpenBSD 8.0 ! considering how AI-slop coding has taken over the Linux kernel, recent Ubuntu derive into consumer-grade and security flaws since the systemd avalanche, OpenBSD is my top choice for mission critical headless servers.
15
29
4
276
16,181
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Replying to @aramh
What I mostly hate is 50gb folders for building a hello world ui, I dont like rust
2
1
7
438
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Replying to @jalololoe
Ada sebuah buku yg menurut saya sangat baik buat orang soshum. Physics for Poets. Isinya fisika tanpa dipusingkan oleh persamaan matematika
3
16
1
308
8,503
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
introducing Orvix AI Plan AI sepuasnya, gak perlu ngitung token lagi 13 model dan bakal nambah terus. cocok buat yg pakai OpenClaw atau Hermes sesekali tanpa pusing kuota juga jalan di OpenCode, Pi, omp, API, dan Orvix One Rp100rb / 30 hari, fair use platform.orvix.id/token-plan
Replying to @algonacci
Pak, tidak ada student plan pak? 100.000 dengan kondisi yang disesuaikan juga boleh.
13
9
1
145
9,533
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Replying to @tobi @fmzakari
all this work started down in mexico at the taco sprint fest. 🌮🌮🌮. crazy how a small collective of community rejects (we had the highest concentration of banned from nixpkgs community per square inch all in the same room) yammering about things to solve can have [0] impact. [0] “we gotta find a way to make omarchy in nix happen. arch is the wrong destination but omarchy is the right idea”
3
2
1
23
1,324
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Replying to @sudoingX
We can, thats why the 20gb 3080, 22gb 2080ti, 48gb 4090 versions exist in china. It's not a technical issue, its a product segmentation one.
12
7
1
2,411
108,816
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Please add this to your website
31
83
6
2,558
136,965
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Beam is now available in OpenCode - 1M Context - Free for a limited time
255
187
106
4,975
380,805
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
IF YOU LET YOUR CLANKER WRITE YOUR RELEASE NOTES SUMMARY LIKE A MOVIE TRAILER NARRATOR I WILL SUCK YOUR EYES OUT OF THEIR SOCKETS AND TIE YOUR OPTICAL NERVES IN A BOW
1
12
442
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
For your weekend reading: Build with Pi Durable on Cloudflare, a walkthrough of how to create a multiplayer chat agent, including subagents, on Durable Objects. earendil.com/posts/pi-durabl…
40
130
25
1,602
110,564
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Testing Deepseek flash for making CAD with mnft. It's good and cheap! Around $0.01 per session turn :D Maybe, I don't need Sonnet for this 🤔
7
4
1
72
4,904
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
> Egonya tinggi > Hanya cara kerja dia yang bener > Orang lain salah semua Anjay tipikal orang ITB banget ya. Mau jadi engineer ataupun jadi menkeu juga kaya gini
1
5
5
57
2,492
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Replying to @bereknyei
Some of this: * Stop doing calls to the nix-daemon during eval * More compact in-memory representation * Arena allocation instead of GC * Multi-threading with fibers * // unions instead of merges * Bunch of tricks with thunks github.com/numtide/iets/blob…
2
1
17
415
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
a lot of people still don't realize Backplane is free and open source. 1. go to github.com/i2cjak/Backplane 2. use ur existing codex/claude/grok install ??? profit
a lot of people still don't realize HeyPCB is free to get started. 1. sign up, claim your username, and get 100% free AI credits 2. explore and download 11,000+ open-source circuit boards 3. fork any design, edit it with AI, and build together in your browser no credit card required. free design exports. no vendor lock-in. we believe hardware should be open and accessible to everyone. we're still early, still improving, and always listening to feedback on how we can make HeyPCB better for everyone. so go build something real! heypcb.ai/world
12
9
2
191
8,357
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
Menuju 50 hari buat IDSW Conference 2026 nih. Keynote spill lanjut 🔥 Untuk sementara ini dulu: Aria Ghora Prabono. Nanti kita spill lagi ya. Siapa beliau? - PhD HUFS, South Korea (applied AI/ML) - Ex ML Engineer @ Nomura Research Institute, Jakarta
3
16
7
82
7,025
૮༼⚆︿⚆༽つ junior gatekeeping engineer retweeted
When I was a junior engineer I unironically DID propose this as a solution to the halting problem! A program objectively halts or doesn't halt for a given input, so take the set of all lookup tables and encode each in a separate TM. Then one of those TMs will just happen to guess the right answer for each TM in the lookup table, and that will be a valid halt detector. This doesn't tell you which TM in the set actually is the halt detector, but at least you know it exists! The problem with this is that any given TM has a finite size, meaning you can only encode a finite lookup table as part of the definition. So then there must exist TMs outside the lookup table, so you don't have a valid halt detector. Notice this means that it is possible to make a halt detector for all TMs under a given size. So while "does an arbitrary TM halt on an empty input" is undecidable, "does an arbitrary TM with fewer than a billion states halt on an empty input" is decidable. And we do have an algorithm for determining if an N-state TM halts: run it and see if it halts after BusyBeaver(N) steps.
The halting problem is a nothingburger. Proof: (1)Every program either halts or doesn't. (2)Store that answer for every possible program in an infinite lookup table. (3)Retrieve the answer in O(1). (4)QED, Turing BTFO.
19
7
3
259
40,512