Hacker News: csunoser

New comment by csunoser in "Microsoft and OpenAI end their exclusive and revenue-sharing deal"

csunoser — Tue, 28 Apr 2026 13:10:18 +0000

Mayhaps! But I think as far as google, anthropic[1] and apple[2] goes, they do use the tpus for training. Ofc v4 and v5 (older generations of tpus) were more specialized for search related embedding workloads and i could see people not using them for training.

[1]: We train and run Claude on a range of AI hardware—AWS Trainium, Google TPUs - April 6th, Anthropic on Google and Broadcom partnership [2]: "[Apple foundation model]... builds on top of JAX and XLA, and allows us to train the models with high efficiency and scalability on various training hardware and cloud platforms, including TPUs and both cloud and on-premise GPUs" - Apple in 2024

New comment by csunoser in "Microsoft and OpenAI end their exclusive and revenue-sharing deal"

csunoser — Mon, 27 Apr 2026 18:17:26 +0000

I am not sure what context Jensen said that. But midjourney uses tpu. Apple uses tpu. They are no other frontier labs that use it, but Google + Anthropic is 2 out of 3 frontier lab so.....

You could reasonably say that "A majority of frontier labs uses TPU to train and serve their model."

New comment by csunoser in "Building FireStriker: Making Civic Tech Free"

csunoser — Fri, 27 Mar 2026 19:24:46 +0000

I am more sympathetic to the OG.

There are many good criticisms against data center. And yet, the water issue always comes up first. Must we spew false/untruthhood just so our political message is catchy? I suppose yes - in times of war/politics, the laws/truths are silent. But it doesn't have to be so here.

New comment by csunoser in "Mistral AI Releases Forge"

csunoser — Tue, 17 Mar 2026 23:47:08 +0000

Huh. I initially thought this is just another finetuning end point. But apparently they are partnering up with customers on the pretraining side as well. But RL as well? Jeez RL env are really hard to get right. Best wishes I guess.

New comment by csunoser in "Ggml.ai joins Hugging Face to ensure the long-term progress of Local AI"

csunoser — Fri, 20 Feb 2026 16:31:28 +0000

I have used both (albeit 2 years ago, and things change really fast). At the time, Candle didn't have 2d conv backprop with strides properly implemented. And getting Burn running libtch backend was just a lot simpler.

I did use candle for wasm based inference for teaching purposes - that was reasonably painless and pretty nice.

New comment by csunoser in "Lyria 3"

csunoser — Wed, 18 Feb 2026 16:21:43 +0000

~~Is anyone getting 404 on the referral link to `gemini.google.com/music` when clicking on try in gemini button?~~

Seems to be working for some folks now.

New comment by csunoser in "We cut our Mongo DB costs by 90% by moving to Hetzner"

csunoser — Thu, 13 Nov 2025 17:46:45 +0000

Even after reading the source, it doesn’t seem like they were hacked? Or if they were, they were not accused of such.

I do think hand rolling your own thing is fraught. But it is very confusing to equate one mother’s complaint to “they have been hacked”.

PS: The people who made their own s3 rans a baby monitor company. News article is about a mother reporting hearing a weird voice from the baby monitour.

New comment by csunoser in "Show HN: Modeling the human body in Rust so I can cmd+click through it"

csunoser — Fri, 10 Oct 2025 20:56:15 +0000

It seems like you think the author wrote this by hand and paid a great deal attention.

What do you think is the chance that claude code wrote the readme?

New comment by csunoser in "Show HN: Modeling the human body in Rust so I can cmd+click through it"

csunoser — Fri, 10 Oct 2025 19:20:24 +0000

I mean, literally not true. There are 7. The problem is that most of the emojis there don't do anything for the content.

Emojis are not the core problem. Mindlessly letting claude do the work and then farm karma on HN is.

New comment by csunoser in "Show HN: Modeling the human body in Rust so I can cmd+click through it"

csunoser — Fri, 10 Oct 2025 18:50:58 +0000

Maybe this is the future. But I dread looking at perfectly formatted yet sterile readme with too many emojis for comfort.

New comment by csunoser in "Rungis: The Market and the City – A day at Europe's largest fresh food market"

csunoser — Tue, 23 Sep 2025 06:21:20 +0000

Sadly, it seems like it is no longer about wholesale. Nothing wrong with that, but a working wholesales market like Rungis and Toyosu Market has a different kind of functional charm to it.

New comment by csunoser in "The Little Book of Linear Algebra"

csunoser — Tue, 02 Sep 2025 17:34:01 +0000

At around 7.4 orthonormal basis and there after, the tex rendering stops working on the github readme preview page.

Instead, it is replaced with a red error box saying: [ Unable to render expression. ]

I wonder if there is an artificial limit for the amount of latex expression that can rendered per page.

New comment by csunoser in "Nvidia DGX Spark"

csunoser — Thu, 28 Aug 2025 01:09:57 +0000

It does say `Experience up to 1 petaFLOP of AI performance at FP4 precision with the NVIDIA Grace Blackwell architecture.` in the features section.

But yeah, this should have been further up.

New comment by csunoser in "Claude Sonnet 4 now supports 1M tokens of context"

csunoser — Tue, 12 Aug 2025 16:30:05 +0000

They don't say it outright. But I think it is not in Claude Code yet.

> We’re also exploring how to bring long context to other Claude products. - Anthropic

That is, any other product that is not Anthropic API tier 4 or Amazon bedrock.

New comment by csunoser in "Anthropic Is Bleeding Out"

csunoser — Fri, 11 Jul 2025 17:26:03 +0000

People love a widget. But the manufacturer loses money at the current price. Manufacturer rises their price for the widget. The resellers are still willing to buy the widget, but is now charging retail consumer more money.

I am very confused why this means Anthropic is bleeding out. The most important thing is that Anthropic has a thing people love. And can raise the price just fine.

Show HN: Dumb STT/diction script for sway-Linux

csunoser — Fri, 04 Jul 2025 21:14:53 +0000

superwhisper is not on linux. And I didn't want to send my voice to Sam Altman. This should give you a minimal voice to speech pipeline if you use sway and have an nvidia gpu handy.

Comments URL: https://news.ycombinator.com/item?id=44467937

Points: 3

# Comments: 0

Show HN: Whatsbuilding – A search engine for real estate developments

csunoser — Sun, 04 May 2025 15:58:03 +0000

Hi HN, I built whatsbuilding to better track and understand real estate development in Austin and other cities in North America.

Having previously worked in real estate development, I found the public data scattered and wanted an easier way to see not just what was being proposed, but also get some context around timelines and how projects compared - what makes one project take 200 days less than a similar one?

This site indexes Austin (and a few other cities) development permits/proposals for searching, and critically, it calculates processing times with percentile comparisons to show how projects stack up against local norms. Heavily inspired by the great judyrecords, I really appreciate projects that "merely" ties together disparate data of obscure origins. Appreciate any feedback!

Comments URL: https://news.ycombinator.com/item?id=43887472

Points: 1

# Comments: 0