Forgot your password?
Close
typodupeerror

Submission + - Artificial Intelligence used to design brand new viruses (bbc.com)

Submitted by fjo3
fjo3 writes: Artificial Intelligence has been used to design brand new viruses that are fully functional and can replicate in the laboratory, say US researchers.

It is the first time whole genomes have been successfully designed by AI.

The resulting 16 novel viruses were created to infect bacteria and pose no threat to people.

The breakthrough has been labelled a "very significant turning point" in science that could unlock a new era for treating disease. But experts have also warned AI-designed viruses raise "urgent" safety and security concerns.

Comment Re: Hold on to your papers... (Score 1) 154

by bill_mcgonigle (#66272338) Attached to: OpenAI's Astra Solved Decades-Old Math Problems For $2,000

> I wonder how many problems they tried before they solved one successfully.

Was gonna mod Insightful, but the replies aren't getting it.

"Solving this problem cost $2000."

That may be true.

"We tried to solve 50 problems at $2000 each and one was successfully solved."

This can also be true which means:

"We spent $100K and we solved one open problem."

which is fiscally, but not akshually, the same as:

"We spent $100K to solve one problem."

If Astra is "all that", OpenAI should have a pricing tier, "no charge unless the problem is solved."

Now *that* would be an impressive vote of confidence and everybody should be excited about it.

Submission + - An Ancient Poop Explosion Could Have Helped Complex Life Evolve (sciencealert.com)

Submitted by alternative_right
alternative_right writes: If the earliest animals hadn't been such prolific poopers, life might not have diversified into all the weird and wonderful forms we see today.

That's the conclusion reached by a new review study published in Trends in Ecology & Evolution.

The team of scientists in Australia and Germany found that the Cambrian explosion â" a period around 540 million years ago when the ancestors of most animal groups arose â" could have been partly set in motion by an explosion of feces, left by those that came before.

Comment bookworm in progress... (Score 1) 48

I see some of these CVE's are fixed in trixie but not bookworm 6.1.177-1 (yet).

Still other CVE's are fixed in bookworm but not yet in trixie. Seems like a few reboots are in the offing.

First time I've seen several fixes in oldstable before stable. It must be a wild week over there. n.b. fixed, not 'not vulnerable'.

Glad the oldstable crew is hard at work!

Comment Re:Stupid law (Score 1) 54

Is anybody aware of open source hardware/software to do the page turning and photograph/OCR the pages?

Google's rig used a puff of air, IIRC.

I asked somebody in the field and they said, "oh, you want a professional rig - those cost $10K and up.

I was like, "no, I have several rare books I'd like to index and this setup is far less complicated than my $249 3D printer." (I already have a good camera.)

I thought I'd find a hackaday or something but my cursory search wasn't fruitful.

Comment Re:Seems like a job for the OS (Score 1) 31

But how long would it take to implement in every supported OS version and then get all the developers to declare a manifest and update their app?

A policy change can be enforced tomorrow (for better or for worse).

The AI bros might even get Samsung to change their mind. Apparently their 'subcontractors' use this technique to scrape all the content they use for training.

Would threatening to cancel a year's RAM order from the chip subsidiary cause a decision to change in the TV subsidiary? I have no idea how dispersed and decentralized Samsung is.

I'd bet it will stick, though, which is likely for the best. Odds are people aren't running Tor exits on their TV's. Maybe there are legitimate apps I'm not aware of.

Comment Flying Saucers (Score 1) 56

by bill_mcgonigle (#66269680) Attached to: New Spinning Drone Hides In Plain Sight

People keep saying it's hard to get a non-blurry picture of them.

Some have suspected somebody has gravity bubble tech which would bend the light.

This thing just spins but isn't itself a disc, though on average it's s disc while in motion.

Post WWII-Germany/Paperclip there were more visible saucers seen over Washington and such but the Germans were developing solid discs. The blurry ones are more over the past few decades as drone tech has become common.

Curious.

Comment Re:PDF (Score 4, Interesting) 84

> I don't know why anyone would send one in any other format.

I'm told by attorneys that Office is nonnegotiable in Law.

Microsoft could probably raise the price to $500/mo and they'd get it.

Should a citizen be required to pay a creepy Redmond company to interact with the Courts?

It"s at least Unconstitutional in New Hampshire, likely other states as well.

Not that Courts themselves care about Unconstitutionality very often. Nor do they personally pay anything for Office. Empathy tends to be low.

But if one wants to make a decent argument that is a good one.

Submission + - A fundamental flaw leaves LLMs strikingly vulnerable to attack (technologyreview.com)

Submitted by joshuark
joshuark writes: It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a team of researchers argue in a paper presented at the International Conference on Machine Learning, a top AI conference, this month. The claim has huge implications for the safety of this technology.

By taking advantage of this flaw, which concerns how LLMs identify who or what is giving them instructions, the researchers were able to make popular LLMs spit out information they had been trained not to provide, such as how to synthesize cocaine and how to sabotage a commercial aircraft’s navigation system.

“There’s a real probability that this is going to be a problem that’s fundamentally unsolvable,” says Charles Ye, an independent researcher and coauthor of the ICML paper.

Companies will typically hire teams of human testers to try to come up with novel attacks that break existing guardrails, a process known as red-teaming. Model makers also use LLM super-hackers (such as OpenAI’s GPT-Red) that find and exploit weaknesses in other models to automate parts of this process. The goal is then to take those attacks and train a new model to resist them and anything that looks like them.

The problem, says Jasmine Cui, another independent researcher and coauthor of the paper, is that the approach amounts to giving the models a list of things they shouldn’t do. But no list is exhaustive. “It’s like watching The Simpsons and they have Bart writing ‘I will not say something inappropriate to my teacher’ a hundred times,” she says. “And he still does things that are pretty crass anyway.”

The ICML paper describes attacks against several of OpenAI’s models, but Cui and Ye say that they have since seen similar results with models made by Anthropic, Alibaba, and DeepSeek.

Cui and her colleagues wanted to find out why an attack like chain-of-thought forgery was so effective. They suspected it had something to do with the mechanism that LLMs use to keep track of where their instructions are coming from.

But what Cui and her colleagues discovered is that LLMs are in fact very bad at keeping track of different roles. In a series of experiments that looked at what was going on inside a handful of different models, the researchers found that LLMs seem to identify the role of a specific chunk of text not by the tags around it but by the style of that text and the words it contains.

The upshot, the researchers claim, is that all an attacker needs to do to hack an LLM is write text that spoofs a certain role. And because roles are a fundamental part of how LLMs work, no amount of training will fully solve the problem.

Ye is worried that nobody is ready for what’s coming. “There’s going to be a huge economic incentive for people to do jailbreaks and prompt injections,” he says. The best defense could be to expect the worst. Organizations shouldn’t trust LLMs, and they should expect that anything done by agents could be unsafe, he says: “That’s not a great solution, but it just might be what we have to do.”

“It’s really incredible that these things are being deployed everywhere to control super-critical systems,” he adds. “There’s been no study of the fundamental science here. We’re all doing it ad hoc.”

Comment Re:that's a dumb take (Score -1) 83

> How many people does it take to run a data center after taking up all this property and all this land that could have been used for something else?

Some medium number of H1B's?

No, the jobs are not going to the citizens of the county where the citizens are subsidizing for all this for their own enslavement.

Who would ever believe such a thing after the Covid lies?

Slashdot Top Deals

You should never bet against anything in science at odds of more than about 10^12 to 1. -- Ernest Rutherford

Close