Hacker News: imranhou

New comment by imranhou in "Geocodio keeps 300M addresses up to date"

imranhou — Mon, 13 Jan 2025 12:12:35 +0000

How much(%) of your data is really just based on OSM?

New comment by imranhou in "Egui – An immediate mode GUI written in Rust"

imranhou — Thu, 26 Dec 2024 13:10:04 +0000

How do you debug these interfaces in a web browser, as can be done in javascript or even c++ web assembly for that matter(https://chromewebstore.google.com/detail/cc++-devtools-suppo...)?

New comment by imranhou in "Ilya Sutskever NeurIPS talk [video]"

imranhou — Sat, 14 Dec 2024 16:20:04 +0000

This is a very interesting point, in some ways the implicit belief is that we just need to get beyond the 700g limitation in terms of scaling LLM models and we would get human intelligence/superintelligence. I admit I didn't really get the body/brain analogy, I would have been better satisfied with a simpler graph of brain weight to intelligence with a scaling barrier of 700g.

New comment by imranhou in "A transformer supply crisis bottlenecks energy projects"

imranhou — Fri, 13 Dec 2024 10:16:33 +0000

A blue whale floats in water, does the LPT?

New comment by imranhou in "Show HN: Llama 3.2 Interpretability with Sparse Autoencoders"

imranhou — Fri, 22 Nov 2024 22:08:29 +0000

Thanks for sharing! It is certainly interesting to me who is not in the mainstream, I'm sure your intended audience understood what you were saying.

New comment by imranhou in "Show HN: Llama 3.2 Interpretability with Sparse Autoencoders"

imranhou — Fri, 22 Nov 2024 09:19:33 +0000

So basically preventing dead latents from occurring and whenever they do occur to possibly reviving them through the use of auxiliary loss term in the loss function? Thanks btw

New comment by imranhou in "Show HN: Llama 3.2 Interpretability with Sparse Autoencoders"

imranhou — Fri, 22 Nov 2024 09:01:58 +0000

Coming from a layman's perspective, a genuine question regarding: "Implements SAE training with auxiliary loss to prevent and revive dead latents, and gradient projection to stabilize training dynamics".

I struggle to understand this phrase "to prevent and revive ", perhaps this is simple speak to those that understand the subject of SAEs, but it feels a bit self contradictory to me, could anyone elaborate?

New comment by imranhou in "Launch HN: Codebuff (YC F24) – CLI tool that writes code for you"

imranhou — Fri, 08 Nov 2024 07:22:44 +0000

What do you think about having codebuff write a parser for javascript? Something that is specifically built to enhance itself that goes beyond the regular parsers and creates a more useful structure of the codebase to be then used for RAG for code writing? This would be double useful as a great demo for your product as well as enhancing your product intrinsically. For example the new parser can not only build the syntax tree but also provide relevant commentary for each method to describe what it does to better pick code context.

New comment by imranhou in "Physical Intelligence's first generalist robotic model"

imranhou — Thu, 31 Oct 2024 22:32:37 +0000

First of all - incredible work. Do you guys plan to integrate frameworks like ROS to help manage this robot?

New comment by imranhou in "Grandmaster-level chess without search"

imranhou — Fri, 18 Oct 2024 00:29:42 +0000

From the page: "We also show that our model outperforms AlphaZero's policy and value networks (without MCTS) and GPT-3.5-turbo-instruct."

Why compare this to GPT-3.5-turbo-instruct? Is that near SOTA in this space?

New comment by imranhou in "Huly – Open-source project management platform"

imranhou — Mon, 14 Oct 2024 05:52:29 +0000

Postman is one example - imagine spending 30 bucks a month on a tool that lets you call APIs.

New comment by imranhou in "OpenAI to Become For-Profit Company"

imranhou — Thu, 26 Sep 2024 15:32:14 +0000

Based on what I've read it is allowed for a non profit to own a for profit asset.

So I'm assuming the game plan here is to adjust the charter of the non profit to basically say we are going to still keep doing "Open AI" (we all know what that means), but through the proceeds it gets by selling chunks of this for-profit entity, so the essence could be the non-profit parent isn't fulfilling its mission by controlling what openai does but how it puts the money to use it gets from openai.

And in this process, Sam gets a chunk (as a payment for growing the assets of the non-profit, like a salary/bonus) and the rest as well....?

New comment by imranhou in "Launch HN: Modern Realty (YC S24) – AI Real Estate Agent for Home Buyers"

imranhou — Wed, 25 Sep 2024 12:05:13 +0000

Is there any data or statistics you could point to? I’m really curious as to how much shift has occurred since the shift by NaR

New comment by imranhou in "Ask HN: Using LLMs for coding, Is there a preferred language/framework?"

imranhou — Tue, 24 Sep 2024 03:21:15 +0000

What tools do you use to get this done? Is it simply chatting into the portal or something like cursor?

Ask HN: Using LLMs for coding, Is there a preferred language/framework?

imranhou — Sun, 22 Sep 2024 22:51:04 +0000

When build a web app or a mobile app, is there such a thing as best suited language or framework that the LLMs like Claude or gpt4o excel at, significantly better than the next best language/framework? I'm curious as to what your experience has been, haven't been able to find much discussions on this topic.

Comments URL: https://news.ycombinator.com/item?id=41620772

Points: 6

# Comments: 8

OpenAI Strawberry Team [video]

imranhou — Fri, 20 Sep 2024 21:39:02 +0000

Article URL: https://www.youtube.com/watch?v=3k89FMJhZ00

Comments URL: https://news.ycombinator.com/item?id=41605728

Points: 1

# Comments: 0

New comment by imranhou in "After private equity takes over hospitals, they are less adept of providing care"

imranhou — Thu, 08 Aug 2024 13:56:24 +0000

Glad there are non profits running some of them

New comment by imranhou in "AI solves International Math Olympiad problems at silver medal level"

imranhou — Fri, 26 Jul 2024 00:26:48 +0000

If the system took 3 days to solve a problem, how different is this approach than a bruteforce attempt at the problem with educated guesses? Thats not reasoning in my mind.

New comment by imranhou in "CrowdStrike debacle provides road map of American vulnerabilities to adversaries"

imranhou — Sat, 20 Jul 2024 16:03:22 +0000

Windows does not support drivers based on rust language, so perhaps may be another 5 years

New comment by imranhou in "Oil demand set to peak by 2029, major supply glut looms, IEA says"

imranhou — Tue, 18 Jun 2024 02:00:08 +0000

There is a difference between Peak Oil "Output" vs Peak Oil "Demand". The article is about demand and not production capacity which won't matter once the demand starts to go downhill... will never be zero but will not be something that can prop up regimes that have sucked the life out of their people for more than half a century.