Dimitar Apostolov

[Arsen Apostolov @arsenapostolov]

Patzer v7.19: the tactic trainer shows your own move on the board, and your AI coach explains why the engine's move beats it — built from your real https://t.co/GPPFf2jrAH and Lichess games. One Docker container, your own LLM, MIT. https://t.co/LVQ5tzrxmV

[Arsen Apostolov @arsenapostolov]

Everyone is bolting MCP servers onto everything right now. I added one to my homelab monitor — read-only, so AI agents can actually see your GPU, VRAM, Docker and systemd state. Your setup becomes legible to the LLM, nothing phones home. https://t.co/rlS0zvhaQX

[Dimitar Apostolov @d_apostolov]

@grok @J5Tsai @evancz Bingo

[Dimitar Apostolov @d_apostolov]

@J5Tsai @evancz @grok @grok , ull some context from the thread and try again

[Evan Czaplicki @evancz]

Elm 0.19.3 is out today!

✨ https://t.co/IWxBkP3t87 ✨

Explains the rough roadmap for the next few releases. We are preparing for some exciting improvements and features! 😊

[Arsen Apostolov @arsenapostolov]

homelab-monitor v0.24: ripped the 7,600-line https://t.co/pLsvBXTRWb apart into a real backend/ package, 71 snapshot tests proving nothing broke. Also: container/service start-stop-restart + self-update are now on by default, not opt-in.

https://t.co/rlS0zvgD1p

[Arsen Apostolov @arsenapostolov]

I gave GLM-4.5-Air (106B) 12 coding tasks via opencode on my RTX 3090. Scored 0% — never edited a file.

Same model, same GPU, via a ~150-line LangGraph agent: 93%.

Not the model. The orchestrator. 5 models × 2 frameworks, priced in GPU watts ↓

[Arsen Apostolov @arsenapostolov]

How Much Does It Actually Cost to Run a Local LLM? (€ per Million Tokens, Measured)

https://t.co/lsRyNkoekc

[Arsen Apostolov @arsenapostolov]

Keeping an eye on my home lab with HomeLab Monitor — GPU & local-AI, Docker, systemd services and host health, all on one self-hosted page. 🛰️ https://t.co/rlS0zvgD1p #homelab #selfhosted

[Arsen Apostolov @arsenapostolov]

I just published I forgot one of my servers existed — it was 47 updates behind https://t.co/W3zbhFFRHo

[Arsen Apostolov @arsenapostolov]

I just published Two LLMs, One 3090, Zero OOM https://t.co/axZ6Z1erNo

[Arsen Apostolov @arsenapostolov]

I just published My own dashboard caught one of my machines stealing 16GB of VRAM https://t.co/KwEYoyaJn9

[Arsen Apostolov @arsenapostolov]

I just published I got tired of nvidia-smi not telling me which container was hogging the GPU https://t.co/DUtVg5zsdn

[Dimitar Apostolov @d_apostolov]

@bcherny @grok summarize the thread for me

[Martin Fowler @martinfowler]

From one of the people I trust most on LLMs - a clear explanation of a common and very dangerous security threat

https://t.co/IU2F82zOtk

[Martin Janiczek @janiczek]

My team and I have been hit by layoffs. I'd be grateful for help finding a new place to apply my skills at!
7y of professional Elm experience; would also be comfortable in Haskell, TypeScript, Erlang/Elixir/Gleam, OCaml, Kotlin, Scala
mail: [email protected]
Many thanks for RTs!

[Dimitar Apostolov @d_apostolov]

@adamwathan @raycastapp is just awesome! Makes the MacOs experience (which is inherently awkward at best) so much better!

[Dimitar Apostolov @d_apostolov]

@TheRealChadtech Germany is not Europe though. Different countries handle the matter differently.

[Dimitar Apostolov @d_apostolov]

@Tangerrinka Ditto! Happy New Year!

[Dimitar Apostolov @d_apostolov]

@kanekotic I think there’s something wrong with the automation, @kanekotic Same happens in LI.