Hacker News: bconsta

New comment by bconsta in "The classifiers Anthropic puts in front of Fable are too zealous"

bconsta — Thu, 09 Jul 2026 01:51:56 +0000

Kind of ironic considering Anthropic hasn't even been around for 7 years.

New comment by bconsta in "Zuckerberg's war on whistleblowers"

bconsta — Sat, 27 Jun 2026 17:48:11 +0000

Too similar to Goldfinger cheating at cards and golf to not make me chuckle.

New comment by bconsta in "When AI Builds Itself: Our progress toward recursive self-improvement"

bconsta — Thu, 04 Jun 2026 21:51:53 +0000

Seems ironic that Claude isn't listed as a contributor to this article.

If was used in writing the article, why not list it? If it wasn't used, that seems to go against Anthropic's whole message.

Obviously readers value human-written content more, but isn't it their interest to attempt to destigmatize llm output as much as possible?

New comment by bconsta in "Sam Altman’s DRAM Deal"

bconsta — Sat, 06 Dec 2025 02:41:25 +0000

David Graeber in three lines

New comment by bconsta in "How to inject knowledge efficiently? Knowledge infusion scaling law for LLMs"

bconsta — Sun, 05 Oct 2025 00:29:21 +0000

I had no idea about this. Thanks for sharing

New comment by bconsta in "How to inject knowledge efficiently? Knowledge infusion scaling law for LLMs"

bconsta — Sat, 04 Oct 2025 20:46:21 +0000

There is a study that gives a rule of thumb of ~2 bits per param for a model's memorization capacity: https://arxiv.org/abs/2404.05405

New comment by bconsta in "iPhone Air"

bconsta — Tue, 09 Sep 2025 20:28:18 +0000

Pixel 9a is probably the closest to what you're describing on the market today.

New comment by bconsta in "Launch HN: Vassar Robotics (YC X25) – $219 robot arm that learns new skills"

bconsta — Tue, 10 Jun 2025 23:44:33 +0000

If I'm not mistaken, the idea is use MCP to let a user-facing LLM make tool calls to a VLA model with actions the user prescribes. He mentions using the LeRobot library in another comment.

New comment by bconsta in "Thoughts on thinking"

bconsta — Sat, 17 May 2025 00:51:21 +0000

pareto adequate

New comment by bconsta in "Tracing the thoughts of a large language model"

bconsta — Fri, 28 Mar 2025 01:35:52 +0000

Following along these lines, I asked chatgpt to come up with a term for 'haboginking a habogink'. It understood this concept of a 'gorbink' and even 'haboginking a gorbink', but failed to articulate what 'gorbinking a gorbink' could mean. It kept sticking with the concept of 'haboginking a gorbink', even when corrected.