Hacker News: wrxd

Multiple commencement speakers booed for AI comments during graduation speeches

wrxd — Mon, 18 May 2026 09:12:10 +0000

Article URL: https://www.nbcnews.com/video/multiple-commencement-speakers-booed-for-ai-comments-during-graduation-speeches-263486021518

Comments URL: https://news.ycombinator.com/item?id=48177107

Points: 2

# Comments: 0

New comment by wrxd in "An AI Hate Wave Is Here"

wrxd — Mon, 18 May 2026 09:11:10 +0000

Have you all seen the recent commencement speeches?

Speaker: AI is awesome, it's the future. Graduates: booo booo booo booo

https://www.nbcnews.com/video/multiple-commencement-speakers...

New comment by wrxd in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

wrxd — Sun, 17 May 2026 22:41:35 +0000

I also like the index feature form https://maki.sh Source code has a lot of structure, using a real parser instead of grepping and reading files can potentially save a lot of tokens

New comment by wrxd in "AI is a technology not a product"

wrxd — Sun, 17 May 2026 17:00:41 +0000

100% I would be happier to have a small model that can run locally capable of searching the web than a stand-alone frontier model

New comment by wrxd in "Clusters become personal (like PCs did)"

wrxd — Sat, 16 May 2026 22:21:50 +0000

I briefly played with k3s before realising than with a single machine I was maintaining a lot of complexity for limited benefits. Then I switched to NixOS, have everything declared in configuration and a much leaner and simpler setup

New comment by wrxd in "I believe there are entire companies right now under AI psychosis"

wrxd — Sat, 16 May 2026 07:30:40 +0000

I’m at a FAANG and we have $300/day token quota. Personally I don’t use that much of it but management is pushing really hard for it. “the quota has been raised for a reason, use it”. Any task: “have you tried working on it with Claude?”. Every meeting “now engineer x and y will show you what he did with AI”.

It’s not all useless but most of the days I think I would be more productive if some processes were streamlined rather than if I had to throw tokens at them and still fail.

Of all the showcases I’ve seen the best are the ones written by people assuming that the token bonanza will not last so they used AI to build tools they wished they had. AI used to build the tool but by no means used by the tool, so if/when token quota gets reduced we still have a functional tool.

New comment by wrxd in "Toxicity on Social Media"

wrxd — Tue, 12 May 2026 17:59:28 +0000

It could. But most people use social media from mobile apps so it doesn’t really help in the grand scheme of things

New comment by wrxd in "Toxicity on Social Media – The Noisy Room"

wrxd — Tue, 12 May 2026 10:28:14 +0000

The claim that this isn't a hard problem to solve seems very optimistic to me.

The tiny minority dominates the feeds because that's how the incentives for algorithmic driven social media are structured. Do we really expect Meta, X, TikTok to anything that could reduce engagement?

Good luck having any of the mainstream social media apps add the banner they propose.

UK firefighters called to one lithium-ion battery fire every five hours

wrxd — Mon, 11 May 2026 08:42:45 +0000

Article URL: https://www.theguardian.com/uk-news/2026/may/11/uk-firefighters-lithium-ion-battery-fires-ebikes

Comments URL: https://news.ycombinator.com/item?id=48092545

Points: 7

# Comments: 1

New comment by wrxd in "Local AI needs to be the norm"

wrxd — Sun, 10 May 2026 21:50:48 +0000

To be fair the author of the post is using the model Apple provides with the OS so it doesn't have any extra binary size

New comment by wrxd in "Local AI needs to be the norm"

wrxd — Sun, 10 May 2026 21:43:17 +0000

The example in the post confirms my theory that for local models to succeed they need to be "good enough", not big enough that they can compete with frontier models.

They need to be able to do a small task well and they need to be able to run reasonably on consumer-class devices. Even better if they can run on mobile phones.

In my experiments with local LLMs I noticed that while increasing the size of the model is nice the real thing that turns a barely useless model into something useful is the ability to use tools. Giving my models the ability to search the web and fetch web pages did way more to solve hallucinations than getting a bigger model. And it doesn't have a training cutoff. Sure, the bigger model is probably better at using tools but I often find the smaller models to be good enough.

Mlx-serve – run LLMs natively on your Mac

wrxd — Sun, 10 May 2026 08:48:21 +0000

Article URL: https://ddalcu.github.io/mlx-serve/

Comments URL: https://news.ycombinator.com/item?id=48082154

Points: 2

# Comments: 0

Forking the Web

wrxd — Sat, 09 May 2026 11:33:07 +0000

Article URL: https://dillo-browser.org/lab/web-fork/

Comments URL: https://news.ycombinator.com/item?id=48074087

Points: 145

# Comments: 146

Pushing Local Models with Focus and Polish

wrxd — Fri, 08 May 2026 23:57:17 +0000

Article URL: https://lucumr.pocoo.org/2026/5/8/local-models/

Comments URL: https://news.ycombinator.com/item?id=48070244

Points: 11

# Comments: 0

New comment by wrxd in "Zuckerberg 'Personally Authorized and Encouraged' Meta's Copyright Infringement"

wrxd — Wed, 06 May 2026 07:25:20 +0000

Nah, it’s short for metastasis. The only apt name for a company that is after growth an any cost

New comment by wrxd in "Accelerating Gemma 4: faster inference with multi-token prediction drafters"

wrxd — Tue, 05 May 2026 21:49:45 +0000

Thank you!

I already played with Gemma4 on oMLX a while ago. When I have some time I'll check if it supports running MTP models and play a bit more

New comment by wrxd in "Google Chrome silently installs a 4 GB AI model on your device without consent"

wrxd — Tue, 05 May 2026 21:45:04 +0000

I'm not a fan of this being downloaded by default. Still, I very much prefer that, if something if Chrome uses a LLM, that's done via a local LLM rather than by via an API call

New comment by wrxd in "Accelerating Gemma 4: faster inference with multi-token prediction drafters"

wrxd — Tue, 05 May 2026 21:36:29 +0000

I'm not sure I understand how this work https://huggingface.co/google/gemma-4-E4B-it-assistant has 78.8M parameters while the standard variant https://huggingface.co/google/gemma-4-E4B-it has 8B parameters.

Is gemma-4-E4B-it-assistant a model I can use stand-alone or a model I need to use in combination with gemma-4-E4B-it?

New comment by wrxd in "New statue in London, attributed to Banksy, of a suited man, blinded by a flag"

wrxd — Sun, 03 May 2026 21:20:58 +0000

Since last summer a lot of flags appeared all over the UK.

https://en.wikipedia.org/wiki/Operation_Raise_the_Colours https://manchestermill.co.uk/the-men-who-raised-the-flags/

Servers with Personality

wrxd — Sat, 02 May 2026 16:52:10 +0000

Article URL: https://caolan.uk/links/servers/

Comments URL: https://news.ycombinator.com/item?id=47988057

Points: 2

# Comments: 0