Hacker News: joaogui1

New comment by joaogui1 in "If America's so rich, how'd it get so sad?"

joaogui1 — Thu, 23 Apr 2026 18:41:33 +0000

I believe their justification is on the first sentence

> That sort of stuff causes pitchforks to rise up in other countries.

(Not that I agree)

New comment by joaogui1 in "Qwen3.6-35B-A3B: Agentic coding power, now open to all"

joaogui1 — Thu, 16 Apr 2026 14:42:25 +0000

3.6 is model number, 35B is total number of parameters, A3B means that only 3B parameters are activated, which has some implications for serving (either in you you shard the model, or you can keep the total params on RAM and only road to VRAM what you need to compute the current token, which will make it slower, but at least it runs)

New comment by joaogui1 in "AI overly affirms users asking for personal advice"

joaogui1 — Sat, 28 Mar 2026 18:01:32 +0000

The models get deprecated after 1-2 years, so reproducibility is pretty hard anyway (but as others pointed out the paper does list the model versions)

Memory and storage shortages may lead to shipping Steam Machines in 2027

joaogui1 — Sat, 07 Mar 2026 13:00:51 +0000

Article URL: https://www.pcgamer.com/hardware/valve-still-hopes-to-ship-steam-machines-in-2026-but-a-delay-into-2027-is-starting-to-look-like-a-real-possibility-memory-and-storage-shortages-have-created-challenges-for-us/

Comments URL: https://news.ycombinator.com/item?id=47287283

Points: 2

# Comments: 0

New comment by joaogui1 in "The highest quality codebase"

joaogui1 — Fri, 12 Dec 2025 00:52:00 +0000

During pre-training the model is learning next-token prediction, which is naturally additive. Even if you added DEL as a token it would still be quite hard to change the data so that it can be used in a mext-token prediction task Hope that helps

New comment by joaogui1 in "Show HN: Gemini Pro 3 imagines the HN front page 10 years from now"

joaogui1 — Tue, 09 Dec 2025 17:19:24 +0000

HN has been used to train LLMs for a while now, I think it was in the Pile even

New comment by joaogui1 in "Gemini 3"

joaogui1 — Tue, 18 Nov 2025 18:45:28 +0000

Probably figured out the exact cause of the bug but not how to solve it

New comment by joaogui1 in "Gemini 3"

joaogui1 — Tue, 18 Nov 2025 16:45:53 +0000

It says Gemini App, not AI Overviews, AI Mode, etc

New comment by joaogui1 in "Why is Zig so cool?"

joaogui1 — Sat, 08 Nov 2025 14:55:05 +0000

Also bizarre that it got to the front page of HN while being so low quality :/

New comment by joaogui1 in "Signs of introspection in large language models"

joaogui1 — Sat, 01 Nov 2025 14:45:00 +0000

Anthropic has amazing scientists and engineers, but when it comes to results that align with the narrative of LLMs being conscious, or intelligent, or similar properties, they tend to blow the results out of proportion

Edit: In my opinion at least, maybe they would say that if models are exhibiting that stuff 20% of the time nowadays then we’re a few years away from that reaching > 50%, or some other argument that I would disagree with probably

New comment by joaogui1 in "Ink deformation"

joaogui1 — Mon, 06 Oct 2025 17:42:44 +0000

It's their lab notes, so it's exploring a general idea, but they're also referencing previous software they've built (like crosscut)

Gemini Robotics 1.5

joaogui1 — Thu, 25 Sep 2025 17:46:14 +0000

Article URL: https://deepmind.google/models/gemini-robotics/

Comments URL: https://news.ycombinator.com/item?id=45376249

Points: 1

# Comments: 0

New comment by joaogui1 in "We're Joining OpenAI"

joaogui1 — Wed, 03 Sep 2025 21:14:27 +0000

Ads on ChatGPT as a way to extract more money from users

New comment by joaogui1 in "From M1 MacBook to Arch Linux: A month-long experiment that became permanenent"

joaogui1 — Sat, 23 Aug 2025 15:13:02 +0000

Not necessarily meaningless, but maybe relative, i.e. a person who generally replaces non-Apple laptops every X years would replace MacBooks every Y years, with Y > X

New comment by joaogui1 in "Gemini 2.5 Deep Think"

joaogui1 — Fri, 01 Aug 2025 16:22:34 +0000

Mixture of Experts isn't using multiple models with different specialties, it's more like a sparsity technique, where you massively increase the number of parameters and use only a subset of the weights in each forward pass.

New comment by joaogui1 in "Gemini with Deep Think achieves gold-medal standard at the IMO"

joaogui1 — Mon, 21 Jul 2025 17:59:15 +0000

It's not a new product/model with that name, they're just saying it's an advanced version of Gemini that's not public atm

New comment by joaogui1 in "Grok 4"

joaogui1 — Fri, 11 Jul 2025 06:57:41 +0000

They used a text-only subset of HLE

New comment by joaogui1 in "End of an Era"

joaogui1 — Tue, 01 Jul 2025 21:38:01 +0000

DS9 the Star Trek series?

New comment by joaogui1 in "Unsupervised Elicitation of Language Models"

joaogui1 — Sat, 14 Jun 2025 20:21:13 +0000

The fact that she's a scientist communicator doesn't imply that she only did the communication part, I think

What Is Programming?

joaogui1 — Sun, 11 May 2025 15:15:21 +0000

Article URL: https://cacm.acm.org/opinion/what-is-programming/

Comments URL: https://news.ycombinator.com/item?id=43954369

Points: 2

# Comments: 0