Hacker News: enoch2090

New comment by enoch2090 in "Obsidian plugin was abused to deploy a remote access trojan"

enoch2090 — Mon, 11 May 2026 07:09:57 +0000

You better delete all third-party applications for they are having full disk access.

New comment by enoch2090 in "Meta Segment Anything Model 3"

enoch2090 — Wed, 26 Nov 2025 02:13:17 +0000

Looking forward to your progress! Just checked the paper and it says the underlying backbone is still DETR. My guess would be that SAM3 uses more video frames during the training process and caused the dilution of sparse engineering-paper-like data.

New comment by enoch2090 in "Launch HN: Onyx (YC W24) – Open-source chat UI"

enoch2090 — Wed, 26 Nov 2025 02:08:23 +0000

GW! How does Onyx differ from Open WebUI and its alike?

New comment by enoch2090 in "Meta Segment Anything Model 3"

enoch2090 — Tue, 25 Nov 2025 15:13:20 +0000

Surprisingly, SAM3 works bad on engineering drawings while SAM2 kinda works, and VLMs like Qwen3-VL works as well

New comment by enoch2090 in "Nvidia DGX Spark: great hardware, early days for the ecosystem"

enoch2090 — Wed, 15 Oct 2025 04:40:46 +0000

Although a bit off the GPU topic, I think Apple's Rosetta is the smoothest binary transition I've ever used.

New comment by enoch2090 in "Ollama Web Search"

enoch2090 — Fri, 26 Sep 2025 02:44:51 +0000

That's the essence of these services, they never explicitly mention the quota, or secretly lowers it at some point.

New comment by enoch2090 in "Neural Graffiti – Liquid Memory Layer for LLMs"

enoch2090 — Tue, 08 Apr 2025 10:11:03 +0000

Played with the demo a bit and I got confused.

1. The chat context is always provided, and that introduces a bit of uncertainty - when the chat history mentioned something the model is always inclined to connect with it.

2. When I tried to set each context to an empty string, the model doesn't show any evidence of remembering concepts. I told it 5 times that I love cats, and when asked about its favorite animal, its output remains "honeybee" and "octopus".

New comment by enoch2090 in "File Pilot: A file explorer built for speed with a modern, robust interface"

enoch2090 — Thu, 20 Feb 2025 09:20:39 +0000

Can't wait to replace my explorer with this, hope the support will come soon

New comment by enoch2090 in "File Pilot: A file explorer built for speed with a modern, robust interface"

enoch2090 — Wed, 19 Feb 2025 10:07:35 +0000

Looks pretty good but all CJK characters are displayed as questionmarks (???) and switching to a CJK native font does no help. So my user folder is now crowded with ???????? which makes it hard to navigate :(

Boundless Socratic Learning with Language Games

enoch2090 — Mon, 16 Dec 2024 02:33:12 +0000

Article URL: https://arxiv.org/abs/2411.16905

Comments URL: https://news.ycombinator.com/item?id=42427497

Points: 1

# Comments: 0

New comment by enoch2090 in "Ask HN: Any way to borrow compute from Apple M1"

enoch2090 — Thu, 18 Jul 2024 10:56:48 +0000

For llama3 just ask him to install ollama and serve the model. Ollama has auto memory management and will free the model when not used, and whenever you make a call to the API (do let your friend know before you do this) ollama will reload the model back to memory again.

Not sure whether there are anything similar for SD though.

New comment by enoch2090 in "Self-Retrieval: Building an information retrieval system with one LLM"

enoch2090 — Sun, 10 Mar 2024 05:55:37 +0000

I wonder how would they calculate the metrics if the result is generated instead of retrieved? Is it likely that the LLM can generate exactly the same output as the desired result?

New comment by enoch2090 in "FastUI: Build Better UIs Faster"

enoch2090 — Sat, 02 Mar 2024 09:33:23 +0000

I think Qt does have WASM compilation options, but I mean generating interfaces given models like Pydantic/Dataclasses/etc.

New comment by enoch2090 in "FastUI: Build Better UIs Faster"

enoch2090 — Sat, 02 Mar 2024 03:50:02 +0000

Streamlit's state management is so painful, feels like constantly writing hacks.

New comment by enoch2090 in "FastUI: Build Better UIs Faster"

enoch2090 — Sat, 02 Mar 2024 03:45:48 +0000

Actually, does Qt have something like FastUI? Been looking for one.

New comment by enoch2090 in "Altman's Ambition"

enoch2090 — Sun, 18 Feb 2024 02:23:36 +0000

The TOC is a total mess

New comment by enoch2090 in "Time Series Forecasting vs Regression: An informal guide"

enoch2090 — Tue, 13 Feb 2024 09:39:04 +0000

Totally makes sense - I'm already getting used to "ChatGPTing" instead of Googling these questions. I guess the future of these articles is that there are always fields that LLMs hallucinate at, fields that are not very common (explain the products of a small brand, pros &cons of their different models).

New comment by enoch2090 in "A search engine in 80 lines of Python"

enoch2090 — Fri, 09 Feb 2024 06:26:42 +0000

BM25 is just too famous in traditional searching and not worth explaining it all over again.

New comment by enoch2090 in "Show HN: Visualize the entropy of a codebase with a 3D force-directed graph"

enoch2090 — Sun, 04 Feb 2024 05:39:29 +0000

This is really cool! We are recently developing a project with heavy C++ and maybe a little Python scripts & wrappers and we are planning for a major refactor. Is it possible to adopt this with a C++ codebase?

New comment by enoch2090 in "Are we at peak vector database?"

enoch2090 — Fri, 26 Jan 2024 15:19:12 +0000

Normally having a lot of choices is a good thing, but here we are facing a dozen of vector dbs with very similar features - to the root it's just some version of ANN implemented in C++/Rust/whatever, the "peak" means there's nothing new. People are flooding into this field not because there's something worth inventing, but more of fear to lag behind and miss the quick money. That's what I feel about vector DBs in Jan, 2024.