Hacker News: Lwerewolf

New comment by Lwerewolf in "Hitachi launches CO2 heat pump water heaters with solar-friendly tariff controls"

Lwerewolf — Thu, 10 Sep 2026 12:19:04 +0000

There are resident systems, and they're of exactly the type listed in the article (i.e. water heaters). Residential... area heating pumps, so to say, are moving towards propane (afaik). I've been looking into getting a propane monoblock air-to-water pump, with an antifreeze loop towards an indoors heat exchanger / water tank.

New comment by Lwerewolf in "Intrusive Linked Lists"

Lwerewolf — Thu, 03 Sep 2026 14:02:09 +0000

I came to this relatively late (2013-ish, windows kernel development, scouring OSDev, etc) so I thought that was always the right name for them. Prior experience was mostly... higher-level langs.

New comment by Lwerewolf in "GLM-5.3-Flash"

Lwerewolf — Wed, 26 Aug 2026 19:12:25 +0000

As others have mentioned, nothing's stopping any other major provider from offering it. Given its popularity, you can guess how that'll develop. So, overall, irrelevant.

New comment by Lwerewolf in "Tim Curry has died"

Lwerewolf — Wed, 26 Aug 2026 19:09:22 +0000

A regular yearly replay is in order for me, too. Already done with SC2... just SC1/WC3/DoW(2)/RA2(mental/flippedmissions/etc)/RA3(uprising)/HOMM3/etc/etc/etc. Ah well.

New comment by Lwerewolf in "GLM-5.3-Flash"

Lwerewolf — Wed, 26 Aug 2026 15:27:40 +0000

The model weights are MIT licensed.

New comment by Lwerewolf in "Apple introduces M6 and M5 Ultra for a big leap in performance and AI compute"

Lwerewolf — Tue, 25 Aug 2026 14:31:25 +0000

Welcome to the club. If you're _really_ competitive in cs2, I'd swap out to a 9800x3d setup, but it's still a maybe. Very little reason to upgrade right now other than to run LLMs.

New comment by Lwerewolf in "New Mac Studio with M5 Max and M5 Ultra"

Lwerewolf — Tue, 25 Aug 2026 14:03:28 +0000

You need latency for token parallelism, not bandwidth. Hence actual RDMA that bypasses the software TCP stack (ROCe or whatever).

New comment by Lwerewolf in "California's new tire efficiency rules could save drivers $1B a year"

Lwerewolf — Tue, 18 Aug 2026 21:04:31 +0000

You just need social stigma on bad tires. It's the single most important thing on your vehicle, for literally all the reasons out there.

New comment by Lwerewolf in "GPT-5.6 Sol Pricing Cut by 50% on OpenRouter"

Lwerewolf — Tue, 18 Aug 2026 20:39:11 +0000

I literally had Fable tell me that I've checked out a nonexistent upstream commit of LMDB - two days ago. Sol called its BS, of course.

New comment by Lwerewolf in "Israel creates fake think tank in likely attempt to dupe AI chatbots"

Lwerewolf — Tue, 18 Aug 2026 11:28:35 +0000

I mean, a big RAG setup with a smart LLM ought to do it, right? Or even just locally provided knowledge database - given the size of LLMs, what's a clone of wikipedia and whatever else you'd need?

New comment by Lwerewolf in "AI isn’t outthinking mathematicians, it’s out-remembering them"

Lwerewolf — Sat, 15 Aug 2026 22:51:57 +0000

Never been easier to make one.

New comment by Lwerewolf in "The AI Demand Bubble"

Lwerewolf — Tue, 04 Aug 2026 19:22:17 +0000

Poolside's stuff (US) is pretty good.

New comment by Lwerewolf in "An Honest Review of AI Programming"

Lwerewolf — Tue, 04 Aug 2026 13:21:53 +0000

You don't have to. Chatgpt-sol-high already does that for me. Could be the extra instructions or the base system prompt or whatever, point is - it already does it.

New comment by Lwerewolf in "DeepSeek V4 Flash on a Single AMD MI300X"

Lwerewolf — Tue, 04 Aug 2026 11:31:35 +0000

The MI350p exists and should run a decent quant (say, the ~96GB antirez mix) well, but you can get two rtx pro 6000s for one of these, or 8x (actually more) r9700 + probably the gear to run them, etc.

Otherwise, you can probably buy one of these second hand from somewhere (SXM A100s are available that way) and run it in an adapter board.

New comment by Lwerewolf in "Boarding China's Last Bus"

Lwerewolf — Mon, 03 Aug 2026 23:06:43 +0000

Almost reads like state policy-induced FOMO.

New comment by Lwerewolf in "DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis"

Lwerewolf — Fri, 31 Jul 2026 16:17:09 +0000

Just tried the preview on my little test codebase and a "check this out and tell me what you think" prompt used over double the tokens of the previous iteration, but it was a lot more eager as well. Kind of reminds me of the new laguna (s 2.1).

New comment by Lwerewolf in "DeepSeek V4 Flash 0731 Intelligence, Performance and Price Analysis"

Lwerewolf — Fri, 31 Jul 2026 14:56:30 +0000

I've had it run to ~400k when debugging "obscure" (to it) sequences. Wouldn't recommend more.

New comment by Lwerewolf in "Laguna S 2.1"

Lwerewolf — Sun, 26 Jul 2026 01:11:37 +0000

Might wanna try it now, seems to have been largely fixed. Check huggingface threads and reddit. Works for me, very memory hungry and PP speed drops off a cliff around 200k context (~30tok/s decode and ~40tok/s pp - like... hope it isn't debugging with big logs), but it brings a fresh perspective alongside ds4-flash. The more, the merrier - I prefer keen eyeballs more than speed anyways.

New comment by Lwerewolf in "Laguna S 2.1"

Lwerewolf — Thu, 23 Jul 2026 19:34:05 +0000

...they'd messed something up in that quant, apparently fixed promptly. ds4 now supports it as well and it's... well, context-limited (<=250k on a 128gb machine) but it positively flies on an m5 max - the 60tok/s decode / ~500tok/s prefill is real.

New comment by Lwerewolf in "Laguna S 2.1"

Lwerewolf — Thu, 23 Jul 2026 13:21:09 +0000

Tool calling in pi completely broke with the updated q4 gguf (spinquant-less). Guessing it'll take some time.