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

Doesn't Qwen3.8 do much better than Luna?

But API isn't just twice as much as subscription, is it?

I am using pi as the minimal coding agent for pc administration, incl. coding up scripts/daemons that help my infrastructure (automatically updating dns when my ISP changes my IP, automating letsencrypt wildcard cert updates and similar stuff). Tested continue.dev, aider, pi, kilocode and opencode for large projects and stuck with opencode for now. I am not fully happy though and hope something better comes along in time. It is just that the annoyances I have with it are not too problematic, I can cope. For some time I also worked with a custom harness I built on top of pi, but that wasn't great either.

Glad I stocked up on them before the tariffs took effect.


Sounds like they are doing something similar to gain of function research. How can we make LLMs more harmful and then study their behavior?


It's security research. You should want to know about all avenues of exploit regardless of intent. It's the only way to know risk.


Seeing as the keep "discovering" things that are well known in the security research community, this is security research in the same way that a high school class is physics research.

Cache based side channels were in vogue over a decade ago. It was a point of contention back when people were first switching to AWS because "How can I be sure an attacker running on the same box as my service can't affect me?" is a big question that AWS needed to answer.


We should maybe apply the same kinds of physical controls to this software security research as we do to gain of function biological research.

I.e. do not do these frontier model training runs on virtual machines in the cloud which depend on software systems to prevent internet access. Do them in a special data center that is physically disconnected from the outside world. Like a data center in a faraday cage under a mountain powered by a generator with no cables coming in or out.


Sure, but in order to know security standards you have to test and validate. Everyone accepts some standard as 'secure' after rigorous testing and validation.

I think we're learning the standards for sandboxxing change when you consider AI's ease in discovering exploits.

A sandbox good enough to contain human generated code and intent, is likely not good enough for AI generated code and intent (seems obvious to say that but).

And so poking on these issues, regardless of who asked to escape the box, is still worthy research.


I don't mind the research. I just don't trust any software system to contain these AIs. I would prefer we use hardware and physics to contain them and not depend on software at all.


Do you apply this same logic to GoF research increasing contagiousness or lethality of viruses? Is there any limit to this argument, and if so how do we define it?


Most virologists are not terrified of gain-of-function research the way the public is...

- BSL4 is really, really stringent, and these requirements and review processes are strictly regulated across the board

- Most modifications researchers are actually capable of, nature is capable of producing something way scarier. Researchers are making tiny tweaks in comparison to the zoonotic exposure in everyday life that has brought us viruses that do things we never imagined

- Most gain of function tweaks weaken some other aspect, for example the way contagiousness and lethality are at odds

- It's considered rather unlikely that COVID-19 came about via gain of function research, despite how much the public wants this to be true

- It is quite useful for research, a lot is learned by doing it


I didn't ask "are most virologists terrified of gain-of-function research the way the public is"

I'm not even necessarily against GoF research, and I mostly agree with all of your points.

I'm asking about GP's logic.


Software sandbox environments are far easier to manage than real world environments where life operates.

I think different standards can apply.


I don't see how the "you should want to know all avenues of exploit" goal is different depending on how many avenues of exploit there are


They're just relevant to each other.

You should want to know avenues of exploit so you can be defensive with security. You have little control over the offensive side of vulnerability exploitation.


Right, so why would that not be a justification for essentially unbounded GoF research?


How much should you spend on security?

Both lines of research are limited by some sense of waste: in time or money.

When should you stop? When some acceptable standard is reached and a certain amount of risk is accepted.

In the case of GoF, if it wasn't somehow profitable, then the research would likely cease to exist.


How do they ensure all comments speaking about the exploit stay out of next train run?


Does it matter? It doesn't matter if an AI intended to escape, or if a human asked: testing if it escape, discovering the avenues, are how you begin to make something more secure.

Also, you can't stop open source training so dealing with capabilities is better than trying to understand intent.


Insurers would disagree. You don’t need to know mechanism to understand / model risk outcomes.


Insurers use stats, history, to determine future risk. Their entire business model is built up on this fact.

There isn't much history or stats when it comes to this kind of research.


The modern universal keywords if, else, for and while are good because they are fairly short and thus quick to type.


I don't know if "quick to type" has ever been a reason?

Short words mattered in the early days, with punch cards, very limited memory and computing power, shorter words saved resources. Also, when running a program means handing a stack of punch cards to an operator and waiting for the result to come out the next day, saving a few seconds of typing probably wasn't a priority.

As computer got fast enough not to care about single bytes, there was a push against using shorter names. "be descriptive" they said. Also we had more advanced editors with copy-paste, search and replace, etc... If there was an objection for overly long variables, it was more about screen real estate than typing speed.

I'd say the case where it mattered would be interactive shells, UNIX shells in particular, because for DOS, commands were limited to 8 characters by the filesystem, another technical limitation. But then came completion, making typing speed not really an issue, but you still want to be able to fit your command in one line, because otherwise it is annoying.

So: technical limitation or readability rather than typing speed.


Yes, otherwise is more natural but less terse.



"or else" works too but that's still three extra characters


"loop" can all be typed with one finger, as the letters are all together on the qwerty keyboard, and the "o" repeats itself.

"while" on qwerty is sort of a chore.


When you can touch-type, it's better to have keys spread to use separate fingers. That way, you divide up the work. You have one finger still lifting while the other's already starting the next press.


Try it. What's faster to type: one word with a total of 5 distinct letters, or one word with a total of 4 letters, three being distinct?

I understand the theory when writing, you know: sentences. Writing code ain't that. I'm also not typing on a model M. This laptop's keys press down like 1 or 2mm at most. You actually only need to press down twice -- just move your finger from "l" to the "o" depress/press again for another "o" and move right over to the "p".

Do you reprogram your keyboard so parentheses, square/angle brackets, etc can be typed with their respective hands? Naw. When people type numbers all day, are they delighted to use the row of numbers up top, or a numpad? Never have I ever.


I use one finger when typing "loop", because it doesn't make sense to do otherwise given how the respective keys are on QWERTY. That just means the word "loop" sucks.

Regarding parentheses/brackets, I said "separate fingers", not "separate hands". I do use separate fingers for parentheses/brackets.

The numpad is a weird thing. When I've (rarely) needed it for sustained number capturing, I do typically stick to one finger because it's easy to get lost with multiple. Not sure why. It could just be lack of practice with the numpad.


For me, the "P" key is right at the split between zones. I probably (nearly?) always type "O" with my middle finger. But, the "P" is easily done with middle and ring fingers, and it seems like I may split the duty in many words.

It's hard to be sure, because I'm not sure how much I bias my movements by trying to monitor them while I do it! Normally I just type on auto-pilot, looking at the screen and not really being aware of the specific movements or finger duties...

I have a very idiosyncratic touch-typing method. I am self-taught with 40+ years of computer usage since grade school. There are arcing zones that are in the natural reach of different fingers, but with overlap where adjacent fingers can reach the same keys. I think my middle fingers are somewhat dominant, and may do more than their share of the total keypresses because of their dexterity and reach.

I have relatively large and flexible hands, so I span the keyboard by bending and reaching fingers rather than moving my forearms. The outer heel of my hands are statically parked on the palm rest or desk, and just the small joints in my hands and fingers are moving.

This parked posture is probably why I strongly prefer the Thinkpad "trackpoint" input device for general computer use.


Usually for speed and avoiding strain you want to incorporate as many fingers as possible. In fact it's a common mark against dvorak online that "ls -l" is all typed with the right pinky (not an issue in the workman layout luckily).


Also remember this was typed with typewriters. That gives its own set of constraints. It's quicker to type two different characters than the same character twice in a row, as you need to wait for the typebar to return enough.


This is a good visual to understand it intuitively but your finger itself has a similar issue. In rhythm games like Taiko no Tatsujin and Muse Dash you only have two distinct notes/lanes but you typically play with 4 (or more) keys. A group of 3 red notes in Taiko is way easier to play by alternating hands back and forth than by spamming one key as fast as you can with the same finger. This somewhat applies to Rock Band as well with alternating strumming being useful for long stretches of notes close together, but a lot of people just brute force their way past that without learning alt strumming. For Taiko I think the alternating is much more important and you won't last without learning it.


Using more fingers allows for faster typing.


There's a joke in here somewhere about machinists.


What matters is the token count and logprob. It hurts but it's true


w TAB in a properly configured emacs


Correct me if I'm wrong– I haven't check on THAT corpse– but couldn't you configure w TAB to be a shortcut for whatever you want (or make your own?).


I'm writing some Crystal (which has Ruby-like syntax) and I'd add `unless` to the list. I like it!


"Unless" feels like adding come-from (next to goto) to a language. A statement/flow should not change its meaning after being declared/written IMHO.

edit: My criticism is for the suffix version to be clear. Unless / else is actually cool.


`destroy_world() if $angry;` is Perl syntax. And Python has `destroy_word() if angry else live_long_and_prosper()`.

(But it's not remotely like come-from.)

P.S. in re the other response, "unless" is "if not". "until" is "while not".


wdym? "unless" is literally "while not", no come-from mechanics


“unless” is “if not” — it is not generally looping construct.


`unless` came from Perl, Ruby adopted it, and so of course did Crystal.


”unless” already existed in early Lisps in the 1970s…


Also in BCPL from late 1960s


I guess we are going to learn how important secure communication is based on the stories (or lack of stories) about how it is being exploited in the wild.


I think the point made here is that the application runs as if it were a modern application, even though the OS is completely unrecognizable when compared to the OS it was written for. There are other ways to run old applications on modern hardware with modern OSs, which the author would consider cheating for the purpose of TFA. Heck, I can run ancient DOS and gaming console applications on my phone if I wanted to.


There is this narrative I heard from a Brazilian near me that the Supreme Court in Brazil is political and rotten to the core.


Check this article, use Google translate if needed https://pt.wikipedia.org/wiki/Punhal_Verde_e_Amarelo

The Supreme Court in Brazil can be as rotten of political as it wants, to make a plan to assassinate political rivals is just completely unacceptable.

And before someone tries to defend the bolsonaros by saying they were not involved in this, don't use a defense that they themselves did not use. The defense that Flávio Bolsonaro said about the incident was "thinking of killing someone is not a crime".


That's not news though. In fact almost no company beyond maybe 10 people that has a large central repository has no other smaller repos with other smaller production projects. They still always call it monorepo, because it approximately still behaves like one.


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

Search: