Hacker News: seahorseemoji

New comment by seahorseemoji in "Muse Wonder"

seahorseemoji — Tue, 13 Oct 2026 02:30:55 +0000

Ah, you mean VendingBench [1]

[1] https://andonlabs.com/evals/vending-bench-2

New comment by seahorseemoji in "MiMo-v2.6-Pro: Intelligence, Performance and Price Analysis"

seahorseemoji — Tue, 22 Sep 2026 10:49:19 +0000

The way Artificial Analysis keeps changing their weights feels kind of like deciding who the winner should be and making the weights reflect that. They’ve been changing their weights to add more weight to improved long-running agentic capabilities, but doing so means they’re reducing the relative importance of world knowledge and of writing ability.

I’ll grant that maybe world knowledge isn’t that important for these models. But writing ability is important for human understanding, and I think the weird turns of phrase and word choices reflect the labs’ underweighting of the importance of human understanding.

New comment by seahorseemoji in "Twitter Viewer – View Twitter Without Account"

seahorseemoji — Wed, 26 Aug 2026 18:42:31 +0000

I wish I understood why the majority of the AI discussions happen on Twitter/X.

New comment by seahorseemoji in "Collaborative Human Agent Protocol (CHAP)"

seahorseemoji — Thu, 20 Aug 2026 02:25:29 +0000

People in the comments are rough! I think this looks cool, nice job!

New comment by seahorseemoji in "Smevals: A small eval suite for evaluating models, prompts, and harnesses"

seahorseemoji — Sun, 02 Aug 2026 01:02:05 +0000

This seems to be a collab between Simon Willison and the guy behind the Superpowers plugin

Smevals: A small eval suite for evaluating models, prompts, and harnesses

seahorseemoji — Sun, 02 Aug 2026 01:02:05 +0000

Article URL: https://simonwillison.net/2026/Jul/31/smevals/

Comments URL: https://news.ycombinator.com/item?id=49140081

Points: 6

# Comments: 1

New comment by seahorseemoji in "The difference between "today's task" and "accretive work""

seahorseemoji — Wed, 08 Jul 2026 12:37:11 +0000

That’s a good point. I am torn between two competing strains of thought, though:

- on the one hand, time-to-market is super important. Getting to the right place faster is obviously better.

- on the other hand, figuring out right product/right fit is hard, and if a business spends that much cost every 3 days chasing every idea (most of which may be bad ideas), they’ve probably wasted a lot of money.

Obviously token costs are cheaper than developers, and local models would reduce costs still further. But the thought I keep coming to is: maybe there’s a benefit to slowing down and not jumping to implement?

I usually hear the opposite side (better to implement 10 things and throw out 9 of them, easier to react to prototypes, etc.). But I also think the infinity of possible ideas doesn’t get smaller when you throw more engineers or compute at it. You just end up exploring more, possibly bad ideas. This works out if exploring more of the space builds a greater understanding of the problem and increases the likelihood that one of your choices pans out. But the cost of exploring the space isn’t $0 and 0 time.

Superpowers 6

seahorseemoji — Tue, 30 Jun 2026 21:28:32 +0000

Article URL: https://blog.fsck.com/2026/06/15/Superpowers-6/

Comments URL: https://news.ycombinator.com/item?id=48739459

Points: 196

# Comments: 77

Ask HN: Anyone else's company ban use of Chinese models?

seahorseemoji — Tue, 23 Jun 2026 13:20:24 +0000

I work for a large company (think banking, insurance, etc.) and we have a ban internally on using any Chinese models. They haven’t stated a reason, but I’m assuming it’s one of the typical ones. I’m wondering how common this is. It feels like a big disadvantage for a company to tie their hands behind their backs like this, given how far ahead the open-source Chinese models are, and given the increasing costs of using the frontier labs’ models.

Comments URL: https://news.ycombinator.com/item?id=48644580

Points: 3

# Comments: 0