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

This is great news for the community. We should measure a bit more closely how tool usage behaves in the wild. The gains you can get from different providers are not marginal.

Sometimes it's not just raw intelligence, but engineering as well


UTF-8 should be a universal tokenizer


Shadow libraries maintainers deserve a Nobel prize for their contributions to humanity. Satoshi would be proud.


Satoshi's pride:

* ability to fund shadow libraries without fear of censorship

* lists with a single item still count as lists


To be fair, the theory with the whole coin thing is solid, and I'd say it should count as something to be proud of even if in reality it gets tainted by speculative investments.


Yeah. I personally think the original bitcoin whitepaper is a work of genius. Balancing the soft game theory incentives with hard cryptography garuntees is really cool.

I'd love to see more systems exploring this combination approach. There is a saying about not being able to solve a social problem with technology. Bitcoin is the blueprint on how to do that.

Its everything that came after that point that is the problem.


> There is a saying about not being able to solve a social problem with technology. Bitcoin is the blueprint on how to do that.

What problem did we solve with bitcoin?


The distributed double spend problem.

I'm not here to argue that its a useful problem to solve, just that the solution is ingenious, and i think the methods used potentially have applications to other problems.



The blockchain isn't the problem. It's supposedly the solution.


You want to stop things happening after other things?


> ability to fund shadow libraries without fear of censorship

Bitcoin is much worse than cash in that regard


sure except for all the reasons cash doesn’t work for this


How would it be feasible for Anna's Archive to accept cash?


That's why most shadow libraries are funded with cash.


who do you hand the cash to in order to fund a website?


Sending cash via snail mail to buy stuff online exists. While Anna's archive does not support it, it certainly exists.


This is also forbidden by several post offices. In my country it's against the law to send money by mail.


Ok, but now you're not comparing bitcoin to cash but rather to cash+mail, which has many more tracking opportunities.


What about Monero?


aaronsw would be proud, too.


the internet's own boy :( I'll be always deeply touched by him


Perhaps he could spare a few coins, chump change to him to help out.


Might need more than a few as the price would tank if his wallets came out of dormancy.


Superman will freak out the world if he kills so he doesnt. Does Satoshi want to avoid freaking out the BC holders?


In this particular case, it would be like sending a cheque to somebody knowing it would bounce.


They made LLM possible, for good or bad.


The name Satoshi Nakamoto literally translates to "central intelligence."


Only when you completely disregard Japanese syntax and the fact that East Asian names tend to be made of Chinese characters with good meanings.



IMO The person making those claims has a very trollish comment history and I suspect they do not actually have deep knowledge of the Japanese language or culture, especially surrounding names and kanji.

The name "Satoshi" exists in kanji form in hundreds of different ways, and as you can imagine, the meanings for those are quite varied. One such common example, 智, does in fact mean intelligence, and can be read as "Satoshi", but there are other examples as well. You can search the ENAMDICT here: http://wwwjdic.biz/cgi-bin/wwwjdic

If you think about this name from the perspective of someone who probably isn't a Japanese expert, but is trying to come up with a believable-sounding name that has a semi-secret meaning like this, I think it makes perfect sense.


I did check the claims in the post and found they're factually correct, within the vagaries of name meanings and etymology. Here's the post:

--

"No, "Satoshi Nakamoto" does not translate to "Central Intelligence" in Japanese. Here's a breakdown of the name: - Satoshi (さとし) is a common Japanese given name, often meaning "wise" or "clear-thinking." - Nakamoto (中本) is a common Japanese surname, with "naka" (中) meaning "middle" or "center," and "moto" (本) meaning "origin" or "foundation."

While "Naka" could be loosely interpreted as "center," and "moto" as "origin," this does not equate to "Central Intelligence." The name does not directly relate to any specific phrase or concept like "Central Intelligence." It's a common Japanese name with meanings unrelated to intelligence agencies or organizations."

--

I checked the commenter's history after your comment and I agree it's often trollish, but this is just ad hominem and nothing to do with evaluating the accuracy of the content of this particular post.

It also sounds unrealistic to me that there are actually hundreds of kanji for "Satoshi", but here there are 130 https://japanese-names.info/first-name/satoshi/, so it's a large number.

What that means is that in fact there are a huge number of possible interpretations, and saying that it "literally translates" to central intelligence is misleading if you don't mention that it "literally translates" to more than a hundred different meanings too.

You could say "one possible loose translation is central intelligence". That would be fair enough.

Here's a discussion of the meaning on stackexchange: https://bitcoin.stackexchange.com/questions/83886/do-%E3%82%...


This is stupid because a UTF8 is a tokenizer that covers all Unicode with a vocab of only 256 (yes, without a K). This is the only way of scaling the bitter lesson with tokenizers. Also, with architectures that span +1M context windows, it’s no longer an argument/issue the reduced context windows.


I'd submit a PR with this idea to improve coverage of agents


Is this Garcon [1], or a new tool?

[1]: https://transformer-circuits.pub/2021/garcon/index.html


Hi, paper author here.

This is a new tool which relies on existing introspection libraries like TransformerLens (which is similar in spirit to Garcon) to build an attribution graph. This graph displays intermediate computational steps the model took to sample a token.

For more details on the method, see this paper: https://transformer-circuits.pub/2025/attribution-graphs/met....

For examples of using it to study Gemma 2, check out the linked notebooks: https://github.com/safety-research/circuit-tracer/blob/main/...)

We also document some findings on Claude 3.5 Haiku here: https://transformer-circuits.pub/2025/attribution-graphs/bio...)


No luck so far with: When does the BB(6) halt?


According to gemma3:27b

BB(6) halts after 1,071,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,001,071,000,000,000,000,000,000,000,000,000,000...

and it's still spitting out lines of 000s after 5 minutes. Either a hallucination or a pretty good joke.


considering the amount of bots in HN, not really that much


mmm... like a whole Harry Potter series


My most sincere love to all shadow libraries out there, you're doing god's work.


They do half of the work (which is a helluva lot)... the other half is done by the volunteers that digitize books.

I was looking at my country's "shelve" and it's so sad to see so many missing titles. I almost wanted to go to my local livrary and digitize sone of them. The old ones that are out of print and imposible to acquire right now...

So much knowledge lost.


To be fair, the authors of the books also contribute quite a bit.


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

Search: