Hacker News: bratao

New comment by bratao in "Qwen3.7-Max: The Agent Frontier"

bratao — Wed, 20 May 2026 12:14:22 +0000

It is super strange that all last (3?) releases they keep comparing older models such as Opus-4.6.

MySQL 9.7.0 LTS Is Now Available

bratao — Fri, 01 May 2026 18:33:39 +0000

Article URL: https://blogs.oracle.com/mysql/mysql-9-7-0-lts-is-now-available-expanded-community-capabilities-and-dynamic-data-masking-for-enterprise

Comments URL: https://news.ycombinator.com/item?id=47978326

Points: 1

# Comments: 0

Claude Opus 4.6 accuracy on BridgeBench hallucination test drops from 83% to 68%

bratao — Sun, 12 Apr 2026 18:53:35 +0000

Article URL: https://twitter.com/bridgemindai/status/2043321284113670594

Comments URL: https://news.ycombinator.com/item?id=47743077

Points: 70

# Comments: 20

New comment by bratao in "Tell HN: Litellm 1.82.7 and 1.82.8 on PyPI are compromised"

bratao — Tue, 24 Mar 2026 13:08:05 +0000

Look like the Founder and CTO account has been compromised. https://github.com/krrishdholakia

New comment by bratao in "I Like GitLab"

bratao — Sat, 24 Jan 2026 12:58:54 +0000

One interesting point of GitLab for me is the self-hosted version, including the AI features (Duo) that can also be self-hosted and you can bring your own OpenAI/Anthropic key.

New comment by bratao in "Stop using MySQL in 2026, it is not true open source"

bratao — Sun, 18 Jan 2026 18:16:44 +0000

From my experience MariaDB is not necessarily better than MySQL. The 8.x line brought many interesting features. I dream on switch to Postgres, and try every year but for my use case MySQL is still superior (100Bi+ rows for large texts, and heavy modified - I´m also space constrained - So I need data compression and the VACUUM are not good.)

The percona distribution is very good!

Tell HN: X/Twitter is now removing the chronological timeline

bratao — Thu, 01 Jan 2026 23:26:58 +0000

I really feel like the platform is actively fighting against my own agency, and I want to know if others are seeing this new behavior.

I am an engineer from Brazil working with Deep Learning and NLP. For years, my strategy to avoid social media addiction was simple: I curated a list of researchers and labs in my field, and I only browsed the "Following" tab. I ignore the "For You" tab completely. I treat it like an RSS feed: I want to see what happened in order, and when I reach the post I saw yesterday, I stop.

Some time ago (I think last year?), they split the timeline into two tabs: "For You" and "Following". It was annoying that they removed the old global setting to default to chronological, but I could live with just clicking "Following" every time I opened the app.

But things got much worse last week.

My "Following" feed started feeling weird. Less informative, more controversial, and out of order. Then I noticed that inside the "Following" tab, there is now a dropdown menu to sort by "Popular" vs "Recent".

It seems the default is now "Popular" (algorithmic). Even if I manually switch it to "Recent" to get my chronological order back, it resets on the next access.

They are effectively forcing the algorithm down our throats. I imagine an algorithmic timeline is much better for ad injection and "doomscrolling" engagement because every refresh gives you something new, whereas a chronological feed has a "end" state.

It is super frustrating. I really want to move away from X, but the signal-to-noise ratio for niche technical content is still too high there to just leave.

I just want to vent my frustration.

Comments URL: https://news.ycombinator.com/item?id=46459376

Points: 9

# Comments: 2

New comment by bratao in "Cloudflare Global Network experiencing issues"

bratao — Tue, 18 Nov 2025 11:39:46 +0000

The danger of Internet centralization in Cloudflare

New comment by bratao in "HipKittens: Fast and furious AMD kernels"

bratao — Fri, 14 Nov 2025 22:38:42 +0000

One thing I don't understand about Nvidia’s valuation is that right now a small number of algorithms have 'won,' such as Transformers. The data is very important. Compared to the past where customized code was much more common, such as modeling code and HPC, the ecosystem was very important and it was almost impossible to implement all CUDA and related code.

Competitors now only need to optimize for a narrow set of algorithms. If a vendor can run vLLM and Transformers efficiently, a massive market becomes available. Consequently, companies like AMD or Huawei should be able to catch up easily. What, then, is Nvidia’s moat? Is InfiniBand enough?"

Aider-desk – Desktop application for Aider AI

bratao — Sun, 02 Nov 2025 20:29:14 +0000

Article URL: https://github.com/hotovo/aider-desk

Comments URL: https://news.ycombinator.com/item?id=45793140

Points: 2

# Comments: 0

RWKV-8 ROSA – An attention-free neurosymbolic LLM

bratao — Sat, 11 Oct 2025 13:06:36 +0000

Article URL: https://twitter.com/BlinkDL_AI/status/1976912771985146184

Comments URL: https://news.ycombinator.com/item?id=45548861

Points: 5

# Comments: 1

Apple Made ICE Agents a Protected Class

bratao — Thu, 09 Oct 2025 11:40:56 +0000

Article URL: https://migrantinsider.com/p/scoop-apple-quietly-made-ice-agents

Comments URL: https://news.ycombinator.com/item?id=45526267

Points: 6

# Comments: 1

New comment by bratao in "DeepSeek-v3.1-Terminus"

bratao — Mon, 22 Sep 2025 12:56:24 +0000

The link is off. This link works https://api-docs.deepseek.com/updates#deepseek-v31-terminus

New comment by bratao in "Do I Need Kubernetes?"

bratao — Sun, 14 Sep 2025 02:56:45 +0000

I use Rancher for a hosted Kubernetes cluster on top of dozens of dedicated servers, and so far, it has been super nice. What are the alternatives for CI/CD for a small team (30)?

Qwen3-Next: Towards Ultimate Training and Inference Efficiency

bratao — Thu, 11 Sep 2025 17:38:40 +0000

Article URL: https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list

Comments URL: https://news.ycombinator.com/item?id=45214130

Points: 16

# Comments: 1

AMA with Z.ai, the Lab Behind GLM Models

bratao — Thu, 28 Aug 2025 19:19:51 +0000

Article URL: https://old.reddit.com/r/LocalLLaMA/comments/1n2ghx4/ama_with_zai_the_lab_behind_glm_models/

Comments URL: https://news.ycombinator.com/item?id=45055942

Points: 3

# Comments: 0

New comment by bratao in "SpaCy: Industrial-Strength Natural Language Processing (NLP) in Python"

bratao — Tue, 26 Aug 2025 23:15:31 +0000

I'm really curious about the history of spaCy. From my PoV: it grew a lot during the pandemic era, hiring a lot of employees. I remember something about raising money for the first time. It was very competitive in NLP tasks. Now it seems that it has scaled back considerably, with a dramatic reduction in employees and a total slowdown of the project. The v4 version looks postponed. It isn't competitive in many tasks anymore (for tasks such as NER, I get better results by fine-tuning a BERT model), and the transformer integration is confusing.

New comment by bratao in "Following Up on the Python JIT"

bratao — Thu, 31 Jul 2025 16:34:52 +0000

I feel sad and disappointed in Microsoft for letting the entire Faster CPython team go. I was a big supporter, always leaving positive comments and sharing news about their work. I'd figure the team paid for itself in goodwill alone. What a letdown, Microsoft. You ought to do better.

Tesla Q2 2025 Update [pdf]

bratao — Thu, 24 Jul 2025 01:18:29 +0000

Article URL: https://www.tesla.com/sites/default/files/downloads/TSLA-Q2-2025-Update.pdf

Hacker News: bratao

New comment by bratao in "Qwen3.7-Max: The Agent Frontier"

MySQL 9.7.0 LTS Is Now Available

Claude Opus 4.6 accuracy on BridgeBench hallucination test drops from 83% to 68%

New comment by bratao in "Tell HN: Litellm 1.82.7 and 1.82.8 on PyPI are compromised"

New comment by bratao in "I Like GitLab"

New comment by bratao in "Stop using MySQL in 2026, it is not true open source"

Tell HN: X/Twitter is now removing the chronological timeline

New comment by bratao in "Cloudflare Global Network experiencing issues"

New comment by bratao in "HipKittens: Fast and furious AMD kernels"

Aider-desk – Desktop application for Aider AI

RWKV-8 ROSA – An attention-free neurosymbolic LLM

Apple Made ICE Agents a Protected Class

New comment by bratao in "DeepSeek-v3.1-Terminus"

New comment by bratao in "Do I Need Kubernetes?"

Qwen3-Next: Towards Ultimate Training and Inference Efficiency

AMA with Z.ai, the Lab Behind GLM Models

New comment by bratao in "SpaCy: Industrial-Strength Natural Language Processing (NLP) in Python"

New comment by bratao in "Following Up on the Python JIT"

Tesla Q2 2025 Update [pdf]

Reflections on 2 years of CPython's JIT Compiler