Sajarin

They have a bit more info on their announcement blog post[0]

Belfort today released the "so far" CIFAR demo, an encrypted implementation of ResNet-20, a popular model for image classification. It outperforms recent SOTA by 3x with a total latency of less than 200ms

Not many details on how they've done this, so I'm a bit skeptical. Fast HE is a holy grail.

Belfort's image classification is built on top of its upcoming GPU library, Cyclops. It comes with several optimizations that make Cyclops extremely fast on Encrypted AI workloads.

Looks like a lead up to an upcoming library release

[0] https://belfortlabs.com/blog/sofar

Just blogged about this here[0] but at least they're not doing the usual canned PR response surrounding this.

Folks are already building on top of it:

thedavidweng/gork-build[1] — rebrand grok→"gork", stripped vendor telemetry, opt-out-only data retention, blocks x.ai auto-update. A "VSCodium-style privacy fork."

DigiGoon/digi-grok-build[2] — "dgrok" multi-provider CLI, builds from source instead of x.ai CDN.

victor-software-house/open-grok[3] — "opened to every provider."

LukaMucko/grok-build[4] — extra_body support for provider-specific request fields.

RapidAI/grok-build-desktop[5] — Tauri desktop GUI client.

mazdak/grok-build[6] — theming (Catppuccin).

thomas9120/grok-build-archival[7] — Windows telemetry-disable script.

saqoah/grok-build[8] — Kotlin MemoryBackend.

[0] https://news.ycombinator.com/item?id=48928913

[1] https://github.com/thedavidweng/gork-build

[2] https://github.com/DigiGoon/digi-grok-build

[3] https://github.com/victor-software-house/open-grok

[4] https://github.com/LukaMucko/grok-build

[5] https://github.com/RapidAI/grok-build-desktop

[6] https://github.com/mazdak/grok-build

[7] https://github.com/thomas9120/grok-build-archival

[8] https://github.com/saqoah/grok-build

What flashcard app are you using? Anki?

You can turn it off, there's a little toggle at the bottom of the page.

There's a writeup here:

https://blog.rice.is/post/doom-over-dns/

People aren't good at detecting AI generated/edited comments, so unsure how effective this policy will be. Though I guess there are still some obvious signs of AI speak like emdashes and sycophantic (it's not X, it's Y!) speech.

Bit of a shameless plug but I wrote a HN AI comment detector game[0] with AI and most of my friends and fellow HN users who tried it out couldn't detect them.

[0] https://psychosis.hn/

[1] https://sajarin.com/blog/psychosis/

Shameless plug but made a similar tree here: https://sajarin.com/blog/modeltree/

Wish this post got more of a response, so thanks so much for giving it a try! Hope it was at least fun (and maybe a bit horrifying) :)

psychosis.hn is a daily game. Every day we fetch three stories from a previous front page of HN, each with 5-7 AI comments threaded into the discussion. They have personas, reply to real people, and sometimes have real comments reparented underneath them.

Thanks for the feedback friend, updated to make it (hopefully) a little easier to read!

Sonnet numbering has been weirder in the past.

Opus 3.5 was scrapped even though Sonnet 3.5 and Haiku 3.5 were released.

Not to mention Sonnet 3.7 (while Opus was still on version 3)

Shameless source: https://sajarin.com/blog/modeltree/

Thanks, that means a lot! Let me know if you have any feedback or suggestions, I would love to work on any improvements :)

Those smooth chunks are all (mostly) public park land. Known as Presidio and part of the Golden Gate National Recreation Area.

I think this comes off a bit too strong (as well as the replies to this to be fair)

The example isn't quite accurate. If a friend bought you lunch, the social norm of reciprocity would incline you towards buying them lunch in the future (i.e part of your paycheck)

Free open source software is a public good. While there is no obligation to give back, giving back helps that public good become more useful to other people (including your future self). I'm against making contribution an obligation, but I'm not against light social pressure upon philanthropists who have the means (which is what the parent comment was doing).

What did Ilya see? (or rather what could he no longer bear to see?)

Academics distorting graphs to make their benchmarks appear more impressive

lavish 1.5 million dollar bonuses for everyone at the company

Releasing an open source model that doesn't even use latent multi head attention in a open source AI world led by Chinese labs

Constantly overhyping models as scary and dangerous to buy time to lobby against competitors and delay product launches

Failing to match that hype as AGI is not yet here

I wonder if anyone has done an analysis on the HN user sentiment on the varying AI models over time. I'd be curious to see what that looks like. Increasingly, I'm seeing more and more people talk positively about Gemini and Google (and having used Gemini recently, I align with that sentiment)

I think Bard (lol) and Gemini got a late start and so lots of folks dismissed it but I feel like they've fully caught up. Definitely excited to see what Gemini 3 vs GPT-5 vs Claude 4 looks like!

Blaming the “system” is easy but is it the whole picture?

How much of it is due to culture? Teachers in western countries are not as respected as teachers in other parts of the world. A few teachers abuse their authority and that results in outrage and lawsuits from parents, rightfully so.

I can imagine in many schools in the US, if a cellphone ban were to be implemented, there would be a large outcry from parents on how restrictive or overreaching that policy would be. Even if the net positives (as shown in the article) are proven to outweigh the pragmatic concerns (i.e I might need to be in communication with my child) why take the risk?

Not to be supporter of “the man” but it seems unfair to point the finger at a system that takes steps to preserve itself without also acknowledging the hostile environment in which it operates.

Parents have greater zeal in suing the school than they have in attending open board meetings.

As a previously precocious young teen, I would have agreed with your point but as someone who is now a boring adult I disagree.

Projects are often and will continue to be judged by their marketing. There are many such cases of "I'm X years old and I made Y" posts on Hacker News reaching the front page. As a founder, you should use whatever you can to get eyeballs on your product. As a hacker, you should try to make something you think is cool.

While it is obviously cool to have something novel or technologically interesting to showcase, the value is often less in the actual product and more in the nostalgia and reminder that we too, as boring adults, were once younger hackers.

Let's not be so hard on each other. I think it's a pretty well designed landing page (although the mobile website needs some work in terms of responsiveness.) I don't think this is similar to Anki because there's no spaced repetition or flashcard retrieval involved (from what I could tell).

It does seem like a tool for cheating which is somewhat questionable. I do like the idea of a young hacker today figuring out how to automate their homework, but I think the tool can be a bit more tailored and more ethical if it focused on a specific use case students would equally pay for (ex: AP test prep)

Yeah, the organizers are definitely taking an extreme route but that's probably the point. Extreme positions get more marketing hype so to speak. Once you have the attention, then you can start to influence a culture of questioning AI and AI usage.

I will say to your first point, it's possible they are trying to prove the model of this type of protest with the biggest and most well known company today to lay the groundwork of other protests at other companies. Someone has to light the match.

On October 21st, 2024 at 12:00 pm Stop AI will peacefully barricade, via sit-in, OpenAI's office entrance gate at 575 Florida St in San Francisco CA.

A longer explanation/commentary here:

1) https://forum.effectivealtruism.org/posts/nDLaCPpZTwK7o2vW9/...

2) https://www.lesswrong.com/posts/bAy6w3spwFrnguJqe/why-stop-a...

I’m curious if you (or anyone else) know of any other services that are paid-only in a traditionally free/freemium product space.

Perhaps for news, or video content, or music, or something else. What else do you use today for free that often feels like a Faustian bargain?

As a random aside, If Google released a paid version of their search engine, would you switch back?

Is everything alright? How were you hacked?

I thought the structure was set up in such a way that only Automattic has commercial rights to the Wordpress trademark. To balance this, Wordpress.org has discretionary power on deciding whether Automattic are good stewards of the trademark. Can Wordpress.org directly ask for the 8%? It seems like it has to be done through Automattic.

This whole situation was handled poorly, but can it really be considered bullying? It definitely speaks to the heart of the eternal problem of open source: the imbalance of givers and takers at the bazaar.

What is the unique value proposition of this book? How does it stand apart from the numerous amount of other books on the same topic?

Marketing is hard when there isn't a clear brand. Branding is hard when you don't have a very simple and clear differentiator to promote.

Isn't the computer science version Peter Novig's, "Latency Numbers Every Programmer Should Know" table?

https://norvig.com/21-days.html#answers

Edit: This version is also quite nice - https://gist.github.com/hellerbarde/2843375

Great caveat, not trying to throw away the baby with the bathwater when it comes to cultural interviews or even personality based assessments. It can be done well. My personal take (which I admit is equally a leap of intuition) is that there are probably more cost-effective ways to gain more signal. Anecdotally for instance, I'm seeing an increase in pair-coding/work-along type interviews where both sides get an intuitive sense of the other's depth of technical expertise as well as a taste of what it's like to work with each other.

Inferring negatives from positive information sharing seems pretty adversarial to me, both in spirit and in substance.

My comment was to question why the interview should be framed as something implicity adversarial to begin with. And why it's probably better to use something like references to validate what it would be like to work with someone rather than rely on our own dubious judgement or the positively biased judgement of the candidate's.

Some of the claims here feel like leaps of intuition. You claim companies don't but many companies, especially ones that are not in the tech industry, do employ personality quizzes in their screening process. Even tech companies, now that I think about it, employ a personality test in the form of a "cultural" interview. I am dubious on how helpful they are but they highlight the real issue which is that there simply isn't enough signal during the interview process.

Every hiring process boils down to a risk mitigation problem and the goal is to get enough positive (or negative) signal to make a hiring decision. If that's the case, making the interview an "adversarial" process seems like an awful idea. As you pointed out, why would a candidate share information that puts them in a bad light? It creates an environment where it is hard to extract meaningful signal.

Furthermore, if there's no free lunch from analyzing personality, then why is it worth thinking about in this way? Perhaps it is more critical and effective to verify their skills and experience with the limited interview time and rely on references to validate their work habits.

It’s a miracle, folks. Donald told the truth for once. It’s the most important election of our lifetimes. And I will win it.

Curiously, his tweet from just 18 hours ago seemed like he was still in the race. I wonder what changed.

Shodan is the villain from System Shock: https://shodan.fandom.com/wiki/SHODAN

SHODAN watches from Security Cameras, stares out of screens and monitors, sends threats and snide messages over the Station's PA system or via email to the player's data reader, and sometimes cuts off communications from friendly sources.