Joined January 2016
- Tweets 60,507
- Following 2,897
- Followers 4,211
- Likes 83,075
1,819 Photos and videos
🦋 hyeseong.kim retweeted
As of today, Hetzner has tripled their prices. Only existing instances are grandfathered in.
Good luck out there. Access to the future will not be free.
my hetzner vpses are now appreciating assets
docs.hetzner.com/general/inf…
16
26
648
63,752
🦋 hyeseong.kim retweeted
Typical coding day with Claude (Opus 4.8)
- explain to Claude the task (5 minutes)
- Claude implements task (10 minutes)
me: "Why is this necessary?"
Claude: "You're right to push back! I over-engineered this!"
- Repeat x87 times (13 hours)
202
243
7,139
594,879
Firefox 152 Now Available With JPEG-XL Support Built By Default, Modernized Settings UI
phoronix.com/news/Firefox-15…
9
77
2,802
🦋 hyeseong.kim retweeted
just got approved @XiaomiMiMo UltraSpeed beta (the fastest model ever) 1T params, ~1,000 tok/s.
it generates faster than i can read. the bottleneck is officially me now 😭
13
10
192
19,512
🦋 hyeseong.kim retweeted
16h
You can now run Kimi K2.7 Code locally! 🌘
We shrank the 1T model to 325GB (-48%) via Dynamic 2-bit where important layers are upcasted.
Run at >40 tok/s on 330GB RAM/VRAM setups.
Run full precision on 610 GB.
Guide: unsloth.ai/docs/models/kimi-…
GGUF: huggingface.co/unsloth/Kimi-…
Jun 12
🌘 Kimi-K2.7-Code, our latest coding model, is now released and open-sourced!
🔷 Improved coding & agent performance over K2.6: 21.8% on Kimi Code Bench v2, 11.0% on Program Bench, and 31.5% on MLS Bench Lite.
🔷 Reasoning efficiency: Less overthinking, with 30% lower reasoning-token usage compared to K2.6.
🔷 Long-horizon coding: Improved instruction following, higher end-to-end coding task success rates.
⚡️ 6x High-Speed Mode coming soon!
🔌 Available today via Kimi API and Kimi Code.
🔗 Kimi Code: kimi.com/code
🔗 API: platform.moonshot.ai
111
211
2,119
620,133
🦋 hyeseong.kim retweeted
BREAKING: Reve 2.0 by @reve debuts at 2nd on Image Editing Arena with an Elo of 1325.
Reve establishes a new Pareto frontier for Preference vs. Speed, faster than any model at this preference level with an average generation time of 86.8 seconds. Reve is now the highest-ranked independent image model company in Image Editing Arena.
Congratulations to the @reve team on this accomplishment!
2
14
64
4,670
🦋 hyeseong.kim retweeted
Formal methods and the future of programming blog.janestreet.com/formal-m…
1
2
250
🦋 hyeseong.kim retweeted
PostgreSQL 19 introduces GROUP BY ALL.
A thoughtful addition that makes aggregate queries cleaner and easier to maintain 👇
#PostgreSQL #SQL
25
248
11,820
🦋 hyeseong.kim retweeted
made a plugin for my blog to render mermaid diagrams with tldraw, with light/dark variants
10
3
232
20,547
🦋 hyeseong.kim retweeted
new post: never waste a token sunilpai.dev/posts/never-was… on using durable objects as a buffer for inference calls
(soon available from cloudflare ai gateway ootb)
14
12
191
18,170
🦋 hyeseong.kim retweeted
React Native 0.86 is now available!
This release includes comprehensive edge-to-edge support on Android 15 . And, following 0.83, it's the second React Native release with no user-facing breaking changes.
reactnative.dev/blog/2026/06…
7
56
403
32,355
🦋 hyeseong.kim retweeted
Jun 13
All 12 modules have been written. I wrote up some thoughts behind this.
kcsrk.info/ocaml/oxcaml/teac…
I'm offering "Functional Programming with OCaml" on the NPTEL platform in July 2026 sem. Enrollment is open now.
The first 8 modules of the interactive book should be fairly stable. The rest is still in development. Sharing early in the spirit of building in the open.
3
12
66
3,531