Hacker News: f4uCL9dNSnQm

New comment by f4uCL9dNSnQm in "New gel restores dental enamel and could revolutionise tooth repair"

f4uCL9dNSnQm — Thu, 06 Nov 2025 11:00:27 +0000

I find it interesting that we are closer to regrowing teeth than hair. At least there was human clinical test for TRG-035. Is it just cause hair transplant is "good enough" stopgap solution?

New comment by f4uCL9dNSnQm in "ChatGPT terms disallow its use in providing legal and medical advice to others"

f4uCL9dNSnQm — Thu, 06 Nov 2025 10:49:24 +0000

Last time I used GPT-4.5 to analyze blood results it gave different output if I uploaded it as 2 instead of 3 separated CSV files. It was both amazing experience: clear and easy to understand statements, and list of most common causes. And terrifying: "What about X?", "You ar absolutely right, there where X results included, disregard everything I wrote above, here is the new analysis".

So for me non-deterministic means unpredictable. Yes, there was nothing random or non-deterministic in that case, I could repeat both scenarios multiple times and get same results again. But the result is affected by something I didn't expect to matter. That damages the trust in tool, no matter how we call it.

New comment by f4uCL9dNSnQm in "The Case That A.I. Is Thinking"

f4uCL9dNSnQm — Tue, 04 Nov 2025 12:21:29 +0000

Do we count Google's search AI overview? Because it is shoved in face of million, every day, and it really is only slight improvement over Clippy.

New comment by f4uCL9dNSnQm in "OpenAI signs $38B cloud computing deal with Amazon"

f4uCL9dNSnQm — Tue, 04 Nov 2025 10:42:52 +0000

OpenAI might actually survive, even if investors lose significant part of their investment. It those those companies that took out loans to invest in "AI" or took overpriced shares as a payment that are getting wiped out.

New comment by f4uCL9dNSnQm in "OpenAI signs $38B cloud computing deal with Amazon"

f4uCL9dNSnQm — Mon, 03 Nov 2025 16:20:16 +0000

It says

> Non-API products may be served on any cloud provider.

I am not sure if Bedrock counts. There are 2 OpenAI models already there: https://aws.amazon.com/blogs/aws/openai-open-weight-models-n...

New comment by f4uCL9dNSnQm in "Google suspended my company's Google cloud account for the third time"

f4uCL9dNSnQm — Mon, 03 Nov 2025 14:42:36 +0000

Isn't it already quite bad? I remember HN post about small company where employees' private accounts got terminated for "due to a prior violation or an association with a previously-terminated Google Play Developer account".

New comment by f4uCL9dNSnQm in "Nvidia to invest up to $1B in AI startup Poolside"

f4uCL9dNSnQm — Mon, 03 Nov 2025 11:28:14 +0000

I wanted to complain how Reuters says close to nothing about Poolside in that news article, but Poolside's own web page is as enigmatic as it gets.

New comment by f4uCL9dNSnQm in "Mathematical proof debunks idea the universe is a computer simulation"

f4uCL9dNSnQm — Mon, 03 Nov 2025 09:54:27 +0000

There is no disagreement about that. But the simulation hypothesis claims that is is more likely that what we perceive as reality is a simulation, run by someone else, than that we are in "real" world.

I think it is worse hypothesis than even Drake's equation - there, at least we know all factors, we just have no idea what their value is and with sample size of 1, the current uncertainly is like 40+ orders of magnitude. That still brings us closer to "is there inteligent life in universe" answer than we ever got with "is this universe a simulation".

New comment by f4uCL9dNSnQm in "UBCO study disproves the simulation hypothesis"

f4uCL9dNSnQm — Mon, 03 Nov 2025 09:33:08 +0000

It is hard to even disprove that we aren't a Boltzmann brain that hallucinated entire reality. Assuming the simulation is perfect(or at lest consistent) the only way to falsify it is to get some impossible estimates for required CPU/memory/storage. I think the whole "if simulations are possible, multiple ones will be created" falls apart when 1 second of running simulation requires several years of compute.

New comment by f4uCL9dNSnQm in "In orbit you have to slow down to speed up"

f4uCL9dNSnQm — Fri, 31 Oct 2025 15:10:04 +0000

There is the small, tiny issue of people commenting just based on title, without even reading the article.

In this case I expected it just links to https://www.youtube.com/watch?v=bcvnfQlz1x4 and didn't even notice in links to Wired.

New comment by f4uCL9dNSnQm in "3D solar tower increases capacity factor 50%, triples solar surface area"

f4uCL9dNSnQm — Thu, 30 Oct 2025 14:29:32 +0000

So how does it compare to simpler vertical bifacial solar panels discussed in https://news.ycombinator.com/item?id=45414215 ?

New comment by f4uCL9dNSnQm in "Grammarly rebrands to 'Superhuman,' launches a new AI assistant"

f4uCL9dNSnQm — Thu, 30 Oct 2025 12:28:02 +0000

This has to be equivalent of "Long Island Iced Tea Corp" renaming to "Long Blockchain Corp."

New comment by f4uCL9dNSnQm in "Grammarly rebrands to 'Superhuman,' launches a new AI assistant"

f4uCL9dNSnQm — Thu, 30 Oct 2025 11:20:13 +0000

I had exact same issues with "Uber".

New comment by f4uCL9dNSnQm in "uBlock Origin Lite in Apple App Store"

f4uCL9dNSnQm — Thu, 30 Oct 2025 11:04:13 +0000

If MotoApps are doing this, it is impressive they can also do that on Samsung phones:

https://community.verizon.com/t5/Mobile-Network-Archive/Game...

New comment by f4uCL9dNSnQm in "ChatGPT's Atlas: The Browser That's Anti-Web"

f4uCL9dNSnQm — Wed, 29 Oct 2025 10:06:16 +0000

Speaking of anti-web:

https://i.postimg.cc/br7F8NLd/chat-GPT.png

I wonder when webmasters will take theirs gloves off and just start feeding AI crawlers with porn and gore.

New comment by f4uCL9dNSnQm in "uBlock Origin Lite in Apple App Store"

f4uCL9dNSnQm — Wed, 29 Oct 2025 09:24:56 +0000

Which model is that? I have a single "Moto" app on mine that is un-removable and that is all bloat I can find.

New comment by f4uCL9dNSnQm in "Nvidia takes $1B stake in Nokia"

f4uCL9dNSnQm — Tue, 28 Oct 2025 16:13:29 +0000

I always forget that Nokia bought out Siemens part of "Nokia Siemens Networks" and it is now just "Nokia networks".

New comment by f4uCL9dNSnQm in "LLMs can get "brain rot""

f4uCL9dNSnQm — Wed, 22 Oct 2025 09:57:43 +0000

I hate it but I still found it interesting that something like it can be done with just CSS (-webkit-linear-gradient)

Edit: I noticed that replacing it with "standard" "linear-gradient" reverses the direction of gradient.

New comment by f4uCL9dNSnQm in "Researchers complete first human trial on viability of enteral ventilation"

f4uCL9dNSnQm — Wed, 22 Oct 2025 09:41:28 +0000

No, the total surface of intestine is up to 40m^2, while the single lung has 100 m^2 for exchange. It might be a supplementary method at best.

New comment by f4uCL9dNSnQm in "Postman which I thought worked locally on my computer, is down"

f4uCL9dNSnQm — Tue, 21 Oct 2025 10:54:50 +0000

And Dropbox is a thin wrapper around rsync. Tools like Postman are convenient. JS scripting allows to extend it easily, like to make request with random test data. Anyway, the previous company I worked for banned it, as soon as it switched from local files to cloud based "collections". Older versions would be ok, but like others mentioned, Postman doesn't like to hear "no" when trying to auto-update itself.