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

You can use both providers in Pi using subscriptions. Be careful with Anthropic though, it’s against their TOS.


Doesn’t using the Anthropic subscription this way get charged as extra usage, and thus effectively billed at API rates?


Not if you run your own system prompt that does not explicitly contain the term "pi coding agent". Perhaps they string-match against some other stuff, too, but as long as you paraphrase the system prompt a bit, you should be fine.


That's hilarious


I use my Anthropic subscription with Pi. I thought they would start charging me API usage at some point, but it never happened. (If it did happen, I would switch to open weights models.)


...funnily enough, a few hours after posting this, it stopped working.


Oh, I stand corrected. It's been a while since I've used their models.

Glad we can still use OpenAI subscription through Pi.


They had plans to do that for CLaude SDK and claude -p usage, but paused them indefinitely.


Though this program does not apply to Fable. Which is why most security researchers have started flocking to Sol.


It reads like an ad.

Secondly these are "just" IDORs, arguably the easiest class of vulnerabilities.

Thirdly it compares to GPT 5.5 and Opus 4.8.

No, we don't have Mythos at home.


>Thirdly it compares to GPT 5.5

mythos is <10% ahead of gpt 5.5 on all benchmarks, which it gains by being several times the size of opus. had it been economical to provide, it would've been released to the public on day one instead of the marketing circus those effective altruism clowns had exhibited. admitting that it costs >1000% to run inference on a <10% better model would've been very damning.


> it costs >1000% to run inference

do you have a source for this claim? i thought LLM providers earn high margins from inference (charged by token). is this no longer the case?


if a $6000000 cabinet can generate 10000/s tokens of Opus but only 1000/s tokens of Mythos, then Mythos costs 1000% to run no matter the markup.

no one has a source, because no one knows closed model parameter counts. we have only heuristics which strongly indicate that Mythos is simply a big fucking model that any other lab could make an equivalent of.


This was just theorised. The leaked OpenAI financials suggest otherwise (because of shady naming of losses)

The only ones who seem to profit are the ones running smaller Chinese models. Even NVIDIA seems to have to "reinvest" their profits into sponsoring companies to buy their cards now.


In my experience, GLM 5.2 is extremely good at finding vulnerabilities, and more importantly, unlike Opus, I've never seen it refuse a command. It genuinely is a very strong model for finding and fixing vulnerabilities.


More importantly, unlike Mythos and Fable, you can actually use GLM 5.2! It's not just marketingware that got its founder in hot water with the government.


> Thirdly it compares to GPT 5.5 and Opus 4.8.

> No, we don't have Mythos at home.

That's still useful. To paraphrase the kids these days, GLM5.2 is in the room with us, today. Mythos is not. And for us in the EU, it's even more complicated, as Mythos might be with us in the room one day, and go poof the next day, on the whims of political entities that we have 0 control over.

Knowing where open, accessible, local models are is important. We know they're behind. But there comes a time when "good enough" is useful. Even if they're "just IDORs" today, and even if they're behind SotA today.

As someone else said above, GLM5.2 (and other models in the same tier like kimi, dsv4, etc) is / are slowly becoming "good enough" to assist in automated repo prepare work (download, install, test, edit, re-test, etc). And that translates in RL traces ready to be trained into the next generations. That might be more important than x% behind on benchmarks.


Technically we don't have Mythos at all? You guys have access. This tells me we have Opus at home (open weights).


Yeah they straight up say that their criteria is narrow and primarily important for their specific use case. Never let rationality cause your pitchfork to be cast away though!


>from our experience, uBO Lite does not seem to be as good as the original non-Lite version

In what way? I've never noticed a difference.


The uBlock Origin Lite FAQ said In general, uBOL will be less effective at dealing with websites using anti-content blocker or minimizing website breakage because many filters can't be converted into DNR rules (see log of conversion for technical details).[1]

[1] https://github.com/uBlockOrigin/uBOL-home/wiki/Frequently-as...


It's not a freebie, it still requires a subscription and burns tokens twice as fast as Opus.



I could be wrong, but I believe the name is in reference to the Divine Rapier, an item in Dota 2, which is very popular among Russian speakers.




Rapira appears to be a direct latinisation of the name of the language.


DigiD itself is government-owned, but its infrastructure is managed by Solvinity (a private company). Not really different from the US gov running half its stack on AWS.


Okay, maybe let's not take the US as a point of comparison.


Fine. Not really different from most governments relying on private suppliers to manage their infrastructure.


This one's pretty bad but there are some preconditions.

Requires a "rewrite" directive with a questionmark in the replacement string, and then a subsequent "set" directive that references a regex capture group (e.g. set $var $1).

Also the POC assumes ASLR is disabled.



Does any distro disable ASLR by default?

If you were to do it by hand, nginx doesn't come to mind as a likely candidate.


Not the person you asked but I am not aware of any that disable ASLR by default, though most default to 1 which only enables ASLR for applications compiled to enable it vs 2 forcing it on or 3 on some distributions that use a hardened kernel. Rather than trusting any assumptions I prefer to run checksec [1] on every OS I touch. It's an old script but works just as well today as it did long ago. One may find that some applications are missing some basic hardening compile time options. The script is not an exhaustive test of all modern hardening options. Example of ASLR being forced on:

    # sysctl kernel.randomize_va_space
    kernel.randomize_va_space = 2
Typical invocation:

    checksec.sh --proc-all
This invocation will list the status of RELRO, Stack Canary, NX/PaX, PIE of all running daemons. My CachyOS installation for example is missing Stack Canaries for all daemons.

    checksec.sh --fortify-proc 732
    * Process name (PID)                         : sshd (732)
    * FORTIFY_SOURCE support available (libc)    : Yes
    * Binary compiled with FORTIFY_SOURCE support: N
Some additional compile time hardening options [2] and discussion [3]. Even Rust apparently has some compile time security related options.

[1] - https://www.trapkit.de/tools/checksec/ # some Linux repositories already contain "checksec".

[2] - https://best.openssf.org/Compiler-Hardening-Guides/Compiler-...

[3] - https://news.ycombinator.com/item?id=43533516


I think "rewrite" is rarely used nowadays? Isn't it something from old days of PHP and Apache?


"old days of PHP and Apache" ...

Apache still runs about 23-28% of websites (with some measurements suggesting it is pretty close to equal with nginx). PHP is still in use by 70-80% of websites (numbers vary depending on where you look).

You make it sound like both pieces of tech are irrelevant. Nothing could be further from the truth.

some quick googled examples (like I said other sites' numbers vary, but you get the general idea):

https://www.wappalyzer.com/technologies/web-servers/ https://kinsta.com/php-market-share/


We are talking about an nginx feature (which is commonly used btw), an Apache feature with a similar name and perhaps even vaguely similar functionality is not relevant.


PHP? You mean that little language behind WordPress?


We're enrolled in the Cyber Verification Program and Claude will happily help me look for vulnerabilities and built POCs demonstrating RCE. But when I point it to a malware sample and ask for analysis it will still refuse any work. It's incredibly frustrating.


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

Search: