Hacker Newsnew | past | comments | ask | show | jobs | submit | more codetrotter's commentslogin

> Would love feedback

When I go to the link, the URL is indicating that it will redirect me to a /hn page after I log in.

I write in my email and get sent a login link. I click the button to complete login. I land on a page that asks me to connect PostgreSQL or another data source.

It’s a super small thing of course, and I bet that when I click the HN submitted link again it will redirect me to the /hn since I am now logged in.

But I thought I’d point this out anyhow. Nitpicking is a tradition in these circles ;)

Edit: Clicked the submission again but it’s asking me to log in rather than seeing I am logged in so another nitpick on that also.


Sorry! I noticed this after I submitted. The link should be https://app.camelai.com/hn/


Make a PDF or whatever kind of document containing the things you want to reveal.

Make a hash sum of the file.

Publish the hash of the document somewhere. Like on your Mastodon timeline or on Reddit, or on HN, or as an ad in a national printed newspaper.

Make sure the text is pretty long though. If you write something super short someone could guess the hash for it.

Write a short comment along with the hash. Something like:

“My prediction will be revealed on Dec 31 2025. Sha256 a9774a1a6ebf564cc408cfd86b5f2c06c13d830e143989714d958d34f325db13.”

When time comes, publish the document.


Another alternative is to encrypt the document with a very strong pass key, and publish only the encrypted document now along with an announcement that the key to decrypt it will be published on such and such date.

If you are on a platform that doesn’t support posting attachments you can use IPFS or similar to host the encrypted file.


Bonus points if it's a platform where you can't delete the posts. Or you number the posts like

> My prediction 683 will be revealed on Dec 31 2025. Sha256 a9774a1a6ebf564cc408cfd86b5f2c06c13d830e143989714d958d34f325db13.

> My prediction 684 will be revealed on Nov 7 2025. Sha256 17147091grhef7wy89908409849rwiruriwu87w980241989171n109288021982.

It would even be better to add it to a blockchain to ensure there is no backdating.


If you write something super short someone could guess the hash for it.

You can skip the pdf and unless the thing is so short/predictable you can guess it outright, nobody is going to 'guess' the hash. E.g. 'I made up a random 64 byte sequence, here's its cryptographic hash, guess the sequence' - that doesn't happen.


If the sentence is too short it might be possible to guess.

It’s not going to consist of random bytes.

Imagine for example that a high profile celebrity accused of murder has successfully defended himself in court and on that day he tweets:

I have something to reveal, but I won’t be telling you until the year 2135. Sha256 73fa194342af6b6e355e129fafd9b19d2a63589b1e3c6e2b5d94a7ca1b3e25f6

If this happens to be a short sentence related to recent events, don’t you think someone will figure out what he wrote pretty soon?


You can just pad or HMAC the short sentence.


Absolutely. Padding it falls under making sure the text is not too short.


For anyone else wondering the GitHub repo is here:

https://github.com/software-mansion/popcorn

Couldn’t find any link to it on the article


In Norway we have a website developed by the Norwegian Digitalisation Agency which has the domain

https://www.norge.no/

(Norge means Norway in Norwegian.)

But in our case that’s a little bit different from the Canadian one.

Norge.no is a guide to digital public services in Norway. The portal presents services from national and local government agencies.

So it’s a place you might go to find where to go for government services when you aren’t sure where to go to find it. And then you are taken to the actual website for what you are looking for, which will be under its own domain instead of on norge.no

Personally I don’t use Norge.no, I just go more or less directly to the websites for the government services that I am interested in or I find them on Google.


My late grandfather used to teach some classes many years ago.

He recounted how one time, after he had graded and handed back some deliveries from the students, one of them came up to him with a confused look on his face.

The student said: In the feedback you wrote D.N.A.S.M. What does that mean?

My grandfather said: It means Do Not Abbreviate So Much.

Your comment reminded me of this :)


Meteor got a lot of attention and hype on HN a few years ago.

Looks at Wikipedia article

Well, 12 or 13 years ago probably actually.


whomst'd've'ly'yaint'nt'ed'ies's'y'es


> we got rejected because of "Mac*" in the name. we pointed them to a dozen others that had it, where it seemingly was no problem. didn't help.

To be fair, it’s understandable for many reasons that they don’t want other companies to use “Mac” in the name of the software.

- Brand dilution

- Losing trademark if Mac becomes a generic word

- It’s also annoying actually with apps that name themselves that way. Just because I’m on a Mac doesn’t mean I need that a whole bunch of my apps start with “Mac” in their name. Likewise for apps that start their name with lowercase i on iOS, and apps that end their name with droid on Android.

Also, for the ones that were allowed anyway, were those already big outside of App Store by chance? Or have they been allowing even new apps that don’t have an existing user base into the App Store with names like that?


> Losing trademark if Mac becomes a generic word

> Also, for the ones that were allowed anyway, were those already big outside of App Store by chance?

These two points contradict each other. If Apple were concerned about losing a trademark, they would have already sued popular apps that have "Mac" in the name.


So you agree that “thats because App Store review is a.) random and b.) they play favorites so the same rules don't apply to everyone”


> Losing trademark if Mac becomes a generic word

This is interesting considering that it was a common word before Apple started. Both the variety of apple called Macintosh and a Mac raincoat (named after Charles Macintosh).

Though it would be really funny if the Beatles were referring to a computer in Penny Lane:

"And the banker never wears a mac in the pouring rain. Very strange"


Trademarks are specific to a kind of product or service. A Mac raincoat does not violate a trademark on Mac computers or software.


How about on an MacBook Pro M2 Max with 64GB RAM? Any recommendations for local models for coding on that?

I tried to run some of the differently sized DeepSeek R1 locally when those had recently come out, but couldn’t manage at the time to run any of them. And I had to download a lot of data to try those. So if you know a specific size of DeepSeek R1 that will work on 64GB RAM on MacBook Pro M2 Max, or another great local LLM for coding on that, that would be super appreciated


I imagine that this in quantized form would fit pretty well and be decent. (Qwen R1 32b[1] or Qwen 3 32b[2])

Specifically the `Q6_K` quant looks solid at ~27gb. That leaves enough headroom on your 64gb Macbook that you can actually load a decent amount of context. (It takes extra VRAM for every token of context you need)

Rough math, based on this[0] calculator is that it's around ~10gb per 32k tokens of context. And that doesn't seem to change based on using a different quant size -- you just have to have enough headroom.

So with 64gb:

- ~25gb for Q6 quant

- 10-20gb for context of 32-64k

That leaves you around 20gb for application memory and _probably_ enough context to actually be useful for larger coding tasks! (It just might be slow, but you can use a smaller quant to get more speed.)

I hope that helps!

0: https://huggingface.co/spaces/NyxKrage/LLM-Model-VRAM-Calcul...

1: https://huggingface.co/bartowski/DeepSeek-R1-Distill-Qwen-32...

2: https://huggingface.co/Qwen/Qwen3-32B-GGUF


I really like Mistral Small 3.1 (I have a 64GB M2 as well). Qwen 3 is worth trying in different sizes too.

I don't know if they'll be good enough for general coding tasks though - I've been spoiled by API access to Claude 3.7 Sonnet and o4-mini and Gemini 2.5 Pro.


How do you determine peak memory usage? Just look at activity monitor?

I've yet to find a good overview of how much memory each model needs for different context lengths (other than back of the envelope #weights * bits). LM Studio warns you if a model will likely not fit, but it's not very exact.


MLX reports peak memory usage at the end of the response. Otherwise I'll use Activity Monitor.


I'm also trusting `get_peak_memory` + some small buffer for now.

Still, it reports accurate peak memory usage for tensors living on GPU, but seems to miss some of the non-Metal overhead, however small (https://github.com/aukejw/mlx_transformers_benchmark/issues/...).


There are plenty of smaller (quantized) models that fit well on your machine! On a M4 with 24GB it’s already possible to comfortably run 8B quantized models.

Im benchmarking runtime and memory usage for a few of them: https://aukejw.github.io/mlx_transformers_benchmark/


Yggdrasil is from Norse mythology. So I’m not surprised that a bunch of different people name software after it.

https://en.wikipedia.org/wiki/Yggdrasil

Just like many might pick cool and important names from Greek mythology for example, like Poseidon.

https://en.wikipedia.org/wiki/Poseidon


Brings to mind the system that rules over the Digital World :)

https://wikimon.net/yggdrasill


wow thanks for the link, the UFO style ring interface floating over a honeycomb is extremely similar to a GUI I designed once, didn't realize Digimon had prior art on it.


Indeed; Yggdrasil is also the name of a set of models of the emission from stellar populations:

https://www.astro.uu.se/~ez/yggdrasil/ReadMe.txt

https://www.astro.uu.se/~ez/yggdrasil/yggdrasil.html


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: