Damar Bowolaksono (@damarbowo) profile picture

Long live and prosper 🖖 | ADHD | Data Enthusiast | Math |

github.com/damarbowo
Joined April 2009
Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)Photo by Damar Bowolaksono (@damarbowo)
Damar Bowolaksono retweeted
Jujur model training dan inference engineering adalah topik yang aku gak paham sama sekali. Jadi kali ini bikin explainer untuk diriku sendiri, dengan narasi yang selalu kita tarik ke arah yang relevan ke bisnis Key takeaways: 🧠 Pinternya AI dari training: pre-training (nebak kata berikutnya dari 45 triliun token), lalu post-training (dikasih contoh, dinilai, diajarin guru) 🎛️ Parameter = tombol mixer, ratusan miliar jumlahnya, dan "arti" nyebar di banyak tombol sekaligus 📚 Lebih banyak parameter biasanya lebih pinter, tapi nggak selalu. Data yang dipilih rapi juga ngaruh ✅ AI paling cepet pinter di kerjaan yang hasilnya bisa dicek objektif, kayak kode yang lulus tes. Jagged intelligence katanya Karpathy 🕵️ Belajar dari model lain (distillation) itu lumrah. Nyontek = pakai model orang tanpa izin. Itu tuduhan Anthropic, ga tau bener apa ngga de. Tapi bisa diliat mocin mulai jarang ngeluarin model2 baru sejak ada anti-distillation 🏢 Fine-tuning + LoRA: perusahaan bisa punya AI khusus tanpa bikin model dari nol
1
34
1
250
7,582
Damar Bowolaksono retweeted
Absolute banger
I don't normally agree with random Redditors, but damn. Well said.
110
377
29
7,378
296,203
Damar Bowolaksono retweeted
Perhaps another underrated part of the Riemann Hypothesis solve is that it was done by someone working for a private company, not a faculty member at a university.
This post nails why we why we should take the Riemann Hypothesis breakthrough very seriously, even beyond math. We’re entering an era where solving problems will look very different from how we solved them in the past. “This is not theory-building in the usual sense. It is something much weirder and more alien. The best human mathematicians know how to build theory, painstakingly, to reach a goal. But no human brain, on its own, is capable of creating a wormhole that requires a 25,000 page long tour through existing infrastructure with no obvious or realistic connection to the goal. But huge credit to Levent here: I think no current AI, on its own, is capable of this either.” It used to be the case that to solve a particular problem - code, life sciences, math, cyber, engineering, and so on - you were ultimately bound by how much information an individual or team could retain and understand about the space. More information or more team members don’t always help, because it gets harder to connect the dots or coordinate amongst everyone. AI changes that completely. For the first time ever we can decouple the human context limits and ability to work through information from our ability to drive productivity or discoveries. This is likely not just a small increase in the kinds of discoveries we can drive, but a qualitative shift in human output. Now every single form of prior art, insight, past research, past decision, or unrelated breakthrough, can be used for the discovery or creation of something new. This is useful for anything from cybersecurity to life sciences research. Agents will be able to correlate entirely unrelated work to a separate project; scan every bit of data in your enterprise for solving a problem; and so on. The solutions may not be as elegant as they were in the past, but we sure will have far more of them. We’re only in the first couple of months of actually understanding the implications of this but it’s clearly going to drive a lot of change.
6
1
25
2,135
Damar Bowolaksono retweeted
Distro ini kelakuannya semena-mena yah kayak Windows mentang-mentang merajai pasar distro Linux desktop.
Canonical wants an all-snap, immutable Ubuntu to become the default desktop. Ubuntu Core Desktop, where the kernel, bootloader, desktop and every app ship as snaps, is now targeted for Ubuntu 28.04 LTS. The stretch goal is to make it the default download in 30.04, and classic Ubuntu stays available. Canonical's VP of engineering Jon Seager at Ubuntu Summit 26.10: "taking the whole of the Linux operating system, and cramming the whole thing into snaps." Daily preview images in GNOME and KDE flavours are due very soon.
2
1
5
457
Damar Bowolaksono retweeted
You can't cram deep mathematical thinking into a weekend, and your daily progress will feel painfully slow. Trust that. You have to put in the unglamorous reps. One hard problem. One lemma. One page of notes. Anything. It needs to happen every single day, not in occasional bursts of motivation. Only then does complexity start feeling natural.
1
11
1
89
3,282
Damar Bowolaksono retweeted
Kenapa yg islam pada marah-marah? Justru aneh kalau perpus ditutup perkara jam sholat, library is a place to study. People can spend hours, even days, in the library. Wdym perpus ditutup saat jam sholat? Kultur perpustakaan di Indonesia bener-bener aneh.
baru tau kalo perpus itb tutup tiap waktu sholat jumat... maksud gw kyk... ga semua orang islam... dan ga semua orang cowo...
51
1,304
71
7,968
116,595
Damar Bowolaksono retweeted
Ada kisah lucu waktu D.N. Aidit ditangkap. Jadi dia sembunyi di lemari & ketahuan. Aidit klo ga salah bilang "saya Menko, kamu tidak berhak menangkap saya". Terus yg nangkap dia blg "Mengko apa? Saiki (Sekarang)". Si tentara dengernya Mengko, yg dlm Bhs. Jawa artinya "Nanti".
61 years ago today, D.N. Aidit was executed.
23
587
33
9,126
327,065
Damar Bowolaksono retweeted
Solusi terbaiq, kita perlu siapin ini selalu:
Lu bakal gimana, baru selesai makan di restoran, lupa bawa HP, kartu ga bawa dompet bawa dan lu nggak punya uang tunai
3
21
1,504
Damar Bowolaksono retweeted
Thank you IDSW for the opportunity! First time speaking in front of such a large audience. I got really nervous and ended up saying things that didn't match what was in my head. Well, definitely need to improve my public speaking skills and mentality!
Audio punya deadline. Packet yg datang telat bisa udh ga berguna. Yoseph Bernandus (RevComm Inc) bahas gimana suara jalan dari udara jadi packet, terus balik jadi suara lagi. Buffer bantu, tapi nambah delay. Trade-off ini yg kerasa tiap call nge-lag. #idsw2026
6
5
42
1,095
Damar Bowolaksono retweeted
Kalau mau subscribe OpenCode Go dan pakai mocin aja yang ada ZDR agreement. Dengan skill dan agent spec yang oke tetep bisa kok ngelakuin apa Claude lakuin. Tentunya lebih murah 😁 Claude mau bayar sampe Max atau pake yang Teams juga ga dapet tuh ZDR kecuali yang enterprise. Tapi sialnya pake Claude enterprise malah ga disubsidi, dari "usage limit" jadi "credits" -_-
Worth it sih asal paham cara pakainya, tapi apakah ga resiko bocor? secara kan data perusahaan 🙄
2
4
3
81
3,777
Damar Bowolaksono retweeted
Replying to @zeeg
show me software with 0 bugs and ill show you software with 0 users
155
710
149
11,888
630,584
Damar Bowolaksono retweeted
Replying to @evanpurnama
Itu si zeeg aneh banget sumpah I amatin the last few months Sentry would've benefited from people doing vibe coding -> cuz the best way to secure the vibe-coded app is through production observability They should've embraced and pushed it. Tapi last few months dia malah bashing vibe coding Feels like Boomers who are working on a Gen Alpha product. The product is right, for the right audience, at the right time. Operatornya masih stuck di jaman Jurrasic
4
8
1
79
3,311
Damar Bowolaksono retweeted
Baru nyadar kemarin setelah ada yang posting kalau dulu bapaknya (boomer) gampang banget dapat posisi lumayan di pekerjaan setelah lulus S1 kampus negeri. Dunia memang sedang bergerak ke arah yang tidak mengenakkan. Boomer dulu gampang nyari kerja. Single bread winner di keluarga pun sudah cukup dan bisa punya rumah. Saya sebagai seorang Gen X mungkin masih agak beruntung. Walau gak segampang boomer, tapi S1 itu masih menjanjikan lah, asal mau usaha dikit. Sekarang tantangan Gen Z ke bawah benar2 sulit. Industrialisasi di Indonesia mandeg kalau gak mau dibilang mati. Bubble startup sudah pecah. Ditambah krisis ekonomi dunia gara2 perang. Tanpa itu pun dunia sudah tidak baik2 saja karena lulusan terlalu berlimpah. Dan ini terjadi bukan hanya di Indonesia. Sepertinya kita memang sedang mengalami masalah struktural. Kenaikan jumlah lulusan tidak sebanding dengan pertumbuhan lapangan kerja.
30
486
46
1,969
244,905
Damar Bowolaksono retweeted
Pertanian beras ini nggak terpantau peningkatan subsidinya dalam 2 tahun terakhir karena lewat beberapa pos. Paling gak subsidi pupuk naik dari 26 T jadi 54 T, anggaran SPHP 6 T, belum lain2. Hasilnya? Harga beras medium naik 2000/kg dalam dua tahun. Absurd.
mentan amran cerita kalo pendapatan petani padi sekarang bisa tembus Rp16 juta sebulan, katanya udah ngalahin gaji PNS baru daftar yang cuma Rp5 jutaan caranya pake metode tanam baru gabungan jajar legowo sama teknologi pot dari amerika dan china, bikin populasi padi naik jadi 1 juta rumpun per hektare modal pupuknya disebut cuma nambah Rp2 juta per hektare, tapi bersihnya bisa Rp60 juta dalam empat bulan iya sih kalo lahannya 1 hektare penuh dan panennya mulus 12 ton tanpa hama 💀 sc : CNN Indonesia
11
20
1
63
4,116
Damar Bowolaksono retweeted
Share ah, nemu di Medsos. Buat yg lagi cari list Beasiswa Beasiswa program ASEAN docs.google.com/spreadsheets… Beasiswa Newzeland docs.google.com/spreadsheets… Beasiswa Australia docs.google.com/spreadsheets…
20
1,921
44
9,926
206,622
Damar Bowolaksono retweeted
Mengubah penamaan file dari "Tidak" menjadi "Non", seperti dari "Peraturan Tidak Berlaku" menjadi "Peraturan Non-Berlaku" dsb. Karena secara pribadi lebih suka seperti ini (sebab selaras dengan konsep negasi pada logika, yang mengganti "tidak" menjadi "non"), dan menghemat huruf.
1
3
732
Damar Bowolaksono retweeted
yang paling saya suka dari departemen hukum pidana di FH UGM adalah dosennya ngajar giliran per materi alih alih satu kelas full misal prof. ed ngajar pengantar hukum pidana, yaudah beliau bakal ngajar materi itu di pertemuan 1-2 untuk semua kelas hukum pidana, kemudian dilanjutin mas akbar ngajar materi selanjutnya, dst jadinya gak ada istilahnya war kelas demi ngejar dosen dan gak ada gap pengetahuan karena semua mahasiswa dapet kesempatan di ajar oleh dosen-dosen yang sama
Departemen Hukum Pidana, Fakultas Hukum UGM. Ngumpulin 10 Dosen full team begini, lebih sulit dari mencari dragon ball.
35
272
106
3,848
140,465
Damar Bowolaksono retweeted
Matkul konsen lingkungan yg harusnya ada di fh ugm 1. Reforma agraria 2. Keadilan gender 3. Ekologi politik
7
28
15
313
7,767
Damar Bowolaksono retweeted
John Lennon bisa bangkit dari kubur saking bangganya
13
6
1
63
5,567
Damar Bowolaksono retweeted
Bon Jovi bisa minder mendengar versi kover yang ini
7
10
3
44
6,180