Hacker News: GardenLetter27

New comment by GardenLetter27 in "Omacom Foundation funding hits $10M"

GardenLetter27 — Tue, 25 Aug 2026 08:08:34 +0000

Limited scope isn't a bad thing.

But both have done great work too - I use gnome-keying and KDE Connect all the time.

New comment by GardenLetter27 in "Omacom Foundation funding hits $10M"

GardenLetter27 — Mon, 24 Aug 2026 12:35:54 +0000

This money is funding Hyprland and QuickShell.

New comment by GardenLetter27 in "Pacing the frontier"

GardenLetter27 — Wed, 29 Jul 2026 07:39:42 +0000

As a Europoor I've seen the devastating consequences of rampant safetyism first hand. Entire industries and energy security destroyed to satisfy the precautionary principle.

We need to accelerate as much as possible and create abundance as fast as we can. Life is short.

New comment by GardenLetter27 in "Every new car sold in the European Union must include a driver monitoring camera"

GardenLetter27 — Tue, 07 Jul 2026 22:57:31 +0000

But the EU still blocks self driving cars which would make this unnecessary.

New comment by GardenLetter27 in "Show HN: Oak – Git alternative designed for agents"

GardenLetter27 — Tue, 23 Jun 2026 10:04:31 +0000

How is downloading stuff remotely the bottleneck here?

New comment by GardenLetter27 in "Good results fine tuning a local LLM like Qwen 3:0.6B to categorize questions"

GardenLetter27 — Mon, 22 Jun 2026 15:38:57 +0000

If you're gonna fine-tune for a closed set classification problem like this, you could just fine-tune BERT and get a faster model with better performance.

New comment by GardenLetter27 in "Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?"

GardenLetter27 — Mon, 15 Jun 2026 18:48:17 +0000

Could the harness not check for a failed tool call and pass it to a small model for correction without clogging up the main context?

New comment by GardenLetter27 in "How LLMs work"

GardenLetter27 — Sat, 06 Jun 2026 08:16:53 +0000

It's not just the architecture but also the data - the decoder only approach lets you train in parallel over blocks of text (no RNN serial waiting), that allows you train on much, much more data.

New comment by GardenLetter27 in "Alberta to hold referendum on whether to remain in Canada"

GardenLetter27 — Fri, 22 May 2026 17:21:09 +0000

What? The UK is doing better economically than Germany right now.

And better than anywhere in Europe for Tech with Deepmind and Ineffable Intelligence there.

It still shares many of the same mistakes as Europe though - e.g. now having to buy Russian oil and gas again instead of using the North Sea oil, not expanding nuclear power, rampant welfare and council housing corruption, etc.

New comment by GardenLetter27 in "Mullvad exit IPs are surprisingly identifying"

GardenLetter27 — Fri, 15 May 2026 13:04:17 +0000

Any chance on port forwarding coming back?

New comment by GardenLetter27 in "The other half of AI safety"

GardenLetter27 — Thu, 14 May 2026 07:48:11 +0000

And what will a human do better? Why will the human care? Who will pay the human?

New comment by GardenLetter27 in "If AI writes your code, why use Python?"

GardenLetter27 — Mon, 11 May 2026 21:19:01 +0000

The LLMs just churns out non-idiomatic slop in any language.

It doesn't matter if the 800-line if statement is able to use pattern matching.

There's been a lot of progress on making coding agents able to solve problems when they can easily evaluate in a closed loop, we desperately need something similar for controlling complexity and using relevant abstractions.

New comment by GardenLetter27 in "A recent experience with ChatGPT 5.5 Pro"

GardenLetter27 — Sat, 09 May 2026 12:07:18 +0000

What? DeepSeekV3 just came out and is incredible for the price. Mythos is also half-released.

New comment by GardenLetter27 in "BP profits more than double as Iran war sends oil prices higher"

GardenLetter27 — Tue, 28 Apr 2026 21:19:12 +0000

And all the oil companies - CRC and Chevron are huge.

New comment by GardenLetter27 in "Claude.ai unavailable and elevated errors on the API"

GardenLetter27 — Tue, 28 Apr 2026 21:17:49 +0000

AWS Bedrock supports prompt caching, just note that if you use the Converse API you need to set the cache points manually.

New comment by GardenLetter27 in "AI's economics don't make sense"

GardenLetter27 — Tue, 28 Apr 2026 18:09:57 +0000

The price of everything will go down. That is the beauty of the free market.

New comment by GardenLetter27 in "Microsoft and OpenAI end their exclusive and revenue-sharing deal"

GardenLetter27 — Mon, 27 Apr 2026 18:24:48 +0000

It's not hype, the demand for inference has grown more this year than expected.

New comment by GardenLetter27 in "Microsoft and OpenAI end their exclusive and revenue-sharing deal"

GardenLetter27 — Mon, 27 Apr 2026 18:23:16 +0000

Hopefully they put ChatGPT on Bedrock now.

New comment by GardenLetter27 in "Kimi K2.6: Advancing open-source coding"

GardenLetter27 — Mon, 20 Apr 2026 18:11:51 +0000

The American models also censor a lot of scientific and political views though.

New comment by GardenLetter27 in "OpenClaw isn't fooling me. I remember MS-DOS"

GardenLetter27 — Mon, 20 Apr 2026 17:59:52 +0000

Conservative thinking isn't responsible.

That's how you end up like Germany still using cash and fax machines for 60+ years.