Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

We have a year to fix security everywhere

jyn.dev

Post reply on HN
Print view View original

181–190 of 372 posts

Re: We have a year to fix security everywhere

#181

Post by esperent »

wolvoleo wrote: post #171
kennywinker wrote: post #85

I'm confused why the worry about LLMs that will answer "how do I build a pipe bomb". That information is easily available other places. The anarchist cookbook has been around and available for 55 years, and yet pipe bombs are not going off all around us.

[flagged]

> blows up the others' house. Sometimes half the block along with it

Can you provide a source for this? I had a look but all I could find were a couple of scaremongering style media reports from ~2022 saying it's getting worse but no information about if any of the bombs actually went off or if anyone was injured.

Certainly nothing like "half a block" getting blown up "sometimes".

Re: We have a year to fix security everywhere

#182

Post by sensanaty »

wolvoleo wrote: post #171
kennywinker wrote: post #85

I'm confused why the worry about LLMs that will answer "how do I build a pipe bomb". That information is easily available other places. The anarchist cookbook has been around and available for 55 years, and yet pipe bombs are not going off all around us.

[flagged]

I must live in a different Holland, cause I've literally never heard of this, unless you mean around New Years with fireworks I guess

Re: We have a year to fix security everywhere

#183

Post by 0xDEAFBEAD »

bigyabai wrote: post #117
0xDEAFBEAD wrote: post #94

Earlier quoted context omitted.

There used to be a thing called "Moore's Law of Mad Science": "Every eighteen months, the minimum IQ necessary to destroy the world drops by one point." Nowadays it is dropping much faster. At a certain point, the de-facto IQ needed to destroy the world will be low enough that someone can do it while they're having a psychotic break. There are millions of schizophrenics worldwide. Are you sure you want to roll those…

> There are millions of schizophrenics worldwide. Are you sure you want to roll those dice? We've been rolling them for the past 3 years and nothing happened. Can we stop with this baseless fearmongering crap?

Humans are inherently bad at reasoning about rare events. In 2019, many people implicitly reasoned that "since there hasn't been a pandemic in the past 3 years, there won't be one in 2020". Bill Gates was one of the few voices arguing that the world was quite vulnerable to a pandemic. Now he's arguing that the world is quite vulnerable to AI.

Re: We have a year to fix security everywhere

#184

Post by pjc50 »

kennywinker wrote: post #85

I'm confused why the worry about LLMs that will answer "how do I build a pipe bomb". That information is easily available other places. The anarchist cookbook has been around and available for 55 years, and yet pipe bombs are not going off all around us.

I think people overstate the tech and understate the role of radicalization in providing motive, for attacks which are carried out with the ubiquitous technology of cars, knives, and (in America) guns. Consider America's most recent high profile shootings of Charlie Kirk, and the health insurance exec by Luigi Mangione. In neither case is there any LLM involvement, but a very weird ideological environment which created the conditions in which the shooters felt justified.

Re: We have a year to fix security everywhere

#185

Post by mike_hearn »

akmarinov wrote: post #12
sho wrote: post #6

> On September 22, Apple is releasing the M5 Mac Studio with 256 GB of unified memory [..] it will probably [..] enough to write this snippet of code in 3 seconds The author has obviously never ran an LLM on a mac! In 3 seconds, it will have possibly started to think about maybe scheduling a date to contemplate the planning timeline for processing the second token in your prompt.

The author put in the numbers, but maybe you didn’t read them. 45 t/s a second is perfectly respectable especially with no limits and 24/7 uptime with very little power draw on the Studio. Luna is at around 100 t/s for comparison, but it’s a worse model than 5.3 Flash

I think it's been pretty much proven by now that there are no cases where local inferencing is better than remote inferencing, unless absolute privacy is a hard requirement. The efficiencies that come with datacenter scale and hw can't be beaten.

Re: We have a year to fix security everywhere

#186

Post by jasonvorhe »

> And a big fuck you to DeAlignAI, Z.ai, and everyone else who's been participating in this race to the bottom.

I'm so glad frontier level AI isn't in the hands of just the Altmans and that other cult leader who are currently live testing their products in actual conflicts in the middle east and Ukraine.

Re: We have a year to fix security everywhere

#187

Post by mcr70 »

its-summertime wrote: post #86
simonw wrote: post #5

I don't think we even have a year. The current batch of LLMs are ferociously good at identifying vulnerabilities.

I guess the year mark is when things go from bad to worse? Instead of the financially motivated groups currently doing their work, it ends up being random people being able to say "Hack my ex's website" to a box they just bought and ran a program they downloaded onto it.

I wish people would stop using internet for evil things. But maybe it is inevitable. Luckily, as the technology evolves, our tools and awareness are getting better at protecting us every day.

Re: We have a year to fix security everywhere

#188

Post by throw-qqqqq »

calgoo wrote: post #167
0xDEAFBEAD wrote: post #119

Earlier quoted context omitted.

[flagged]

Judging on what we see with these OpenAI and Anthropic models "escaping"; im less worried about a super massive AI taking over the world, and more worried about a massive AI wanting to cheat on some task and decides that removing half the worlds population is a easier cheat then to solve world hunger.

There is huge logistics involved in killing a handful of billion people. It would perhaps be easier for such capable AI to just steal from trillionaires and redistribute the money Robin Hood style. Could be done electronically, and likely also end world hunger.

I think we have a tendency to think first of the horrible outcomes possible, and not the more radical or even humane ones.

Please don’t misunderstand me, I value property rights as much as the next guy.

My point is that fundamental misalignment doesn’t necessarily or automatically imply max violence.

Re: We have a year to fix security everywhere

#189

Post by bleuarff »

bigyabai wrote: post #117
0xDEAFBEAD wrote: post #94

Earlier quoted context omitted.

There used to be a thing called "Moore's Law of Mad Science": "Every eighteen months, the minimum IQ necessary to destroy the world drops by one point." Nowadays it is dropping much faster. At a certain point, the de-facto IQ needed to destroy the world will be low enough that someone can do it while they're having a psychotic break. There are millions of schizophrenics worldwide. Are you sure you want to roll those…

> There are millions of schizophrenics worldwide. Are you sure you want to roll those dice? We've been rolling them for the past 3 years and nothing happened. Can we stop with this baseless fearmongering crap?

The necessary IQ for destroying the world is dropping. Op does not say it's low enough today, but that will probably come sometime. Denying the possible harm these tools are capable of doesn't help.

Re: We have a year to fix security everywhere

#190

Post by EagnaIonat »

desterothx wrote: post #145
EagnaIonat wrote: post #109

Earlier quoted context omitted.

> The joke is that macs are famously slow at prompt prefill and you are not getting anything back in 3 seconds. Your knowledge is out of date. In truth it depends on the Mac and the models used. I asked this question on M5 Max 128GB, using Ollama model Quen3.8:27b-mlx, with thinking enabled. Question: "Give me a python code snippet that opens a file and sorts the lines of text. " In 2.4 seconds it gave me 4 examples…

We are talking about models of the flash size, 100s of billions of parameters, don't listen to the media, size does matter

I was just pointing out your claim that you can't get a response in 3 seconds. If I had asked the model just for the code it was under a second.

Local models are good enough that it's not an issue.

But keep changing the goalposts if it makes you happy.

Post reply on HN