Hacker News: stri8ed

New comment by stri8ed in "Gemini 3.5 Flash"

stri8ed — Tue, 19 May 2026 18:10:33 +0000

Output cost is 3x from Gemini 3 flash.

New comment by stri8ed in "Gemini 3.5 Flash"

stri8ed — Tue, 19 May 2026 18:09:15 +0000

Given the cost increase associated with this model, and previous model releases, I think the size is trending upwards, not down.

New comment by stri8ed in "An investigation into Substrate"

stri8ed — Sat, 01 Nov 2025 17:03:31 +0000

Is it fair to characterize a founder of a failed startup as a con man? I.e. did he make claims that were factually untrue, and intentionally deceived investors?

New comment by stri8ed in "Flock's gunshot detection microphones will start listening for human voices"

stri8ed — Sat, 04 Oct 2025 17:31:42 +0000

What is the counterfactual? Without knowing the number of attacks prevented by these tools, we don't know what the baseline would be.

New comment by stri8ed in "Flock's gunshot detection microphones will start listening for human voices"

stri8ed — Sat, 04 Oct 2025 17:29:28 +0000

Having adequate law enforcement training and funding, is not mutually exclusive with leveraging technology for more effective enforcement. In fact that's where some of the funding goes. I would be interested in seeing some data reflecting a reduction in crime as a result of increasing the welfare system, as you claim.

https://en.wikipedia.org/wiki/Flock_Safety#Efficacy

New comment by stri8ed in "SWE-Bench Pro"

stri8ed — Mon, 22 Sep 2025 17:18:00 +0000

Not a chance. Even if American companies did abide by it, there is no reason Chinese companies would. And good luck definitely proving that a model trained on it.

New comment by stri8ed in "GPT-5"

stri8ed — Thu, 07 Aug 2025 16:57:05 +0000

That problem along with its many solutions are surely littered throughout the training data. Not to mention, it would be trivial to overfit on that problem. I don't know why people still reference that.

New comment by stri8ed in "Grok 4 Launch [video]"

stri8ed — Thu, 10 Jul 2025 17:21:23 +0000

It's a result of the system prompt, not the base model itself. Arguably, this just demonstrates that the model is very steerable, which is a good thing.

New comment by stri8ed in "OpenAI’s board, paraphrased: ‘All we need is unimaginable sums of money’"

stri8ed — Sun, 29 Dec 2024 23:46:30 +0000

Trained at least in part on Chat-GPT data.

New comment by stri8ed in "OpenAI O3 breakthrough high score on ARC-AGI-PUB"

stri8ed — Sat, 21 Dec 2024 17:14:19 +0000

It would happen in China regardless what is done here. Removing billionaires does not fix this. The ship has sailed.

New comment by stri8ed in "Large Enough"

stri8ed — Wed, 24 Jul 2024 16:25:13 +0000

The benchmarks agree as well.

New comment by stri8ed in "Implementation of Google's Griffin Architecture – RNN LLM"

stri8ed — Thu, 11 Apr 2024 16:51:27 +0000

Isn't that how previous models were, before the attention is all you need paper?

New comment by stri8ed in "DBRX: A new open LLM"

stri8ed — Wed, 27 Mar 2024 17:08:45 +0000

Or be in the business of building infrastructure for AI inference.

New comment by stri8ed in "Ask HN: If you've used GPT-4-Turbo and Claude Opus, which do you prefer?"

stri8ed — Sun, 17 Mar 2024 16:51:58 +0000

You can modify this, by setting the GPT4 system prompt, with instructions of your preferred response style.

New comment by stri8ed in "Ask HN: If you've used GPT-4-Turbo and Claude Opus, which do you prefer?"

stri8ed — Sun, 17 Mar 2024 16:49:59 +0000

For programming, GPT4+. I was excited to switch to Claude after hearing all the positive anecdotes. Having tried it, I'm very unimpressed. It spouted complete, confident sounding nonsense, when I prompted it with a bug I was trying to solve. GPT4, did not get it right initially, but it was more suggestive, instead of wrongly declaring the fault, and lead me to the answer after a few more prompts. Will not be renewing my Claude subscription.

New comment by stri8ed in "A seafood firm wants to farm octopus. Activists say they're too smart for that"

stri8ed — Fri, 09 Feb 2024 00:51:53 +0000

It's impossible to prove.

New comment by stri8ed in "23andMe's Fall"

stri8ed — Wed, 31 Jan 2024 18:21:23 +0000

I don't think you can infer someone's age and skin colon from that post.

New comment by stri8ed in "Noisy brain may underlie some of autism's sensory features"

stri8ed — Fri, 19 Jan 2024 00:37:42 +0000

In humans, a software problem can become a hardware problem, due to plasticity in the hardware. See alcoholism for example.

New comment by stri8ed in "Wikidata, with 12B facts, can ground LLMs to improve their factuality"

stri8ed — Fri, 17 Nov 2023 17:33:29 +0000

Do existing LLM's not already train on this data?

New comment by stri8ed in "Decoding the ACL Paper: Gzip and KNN Rival Bert in Text Classification"

stri8ed — Thu, 20 Jul 2023 23:38:34 +0000

In such a scheme, wouldn't synonyms of the same word be no closer to each other, than any other random string?