cyber risk from frontier AI
“The world is playing with fire on cyber; only inaction has saved us so far.”
26 people · 18 agree · 2 disagree · 5 no clear side · 5 unclear · 66 views · 2026-10-12 to 2026-10-23
Who stands where on the statement above, from their own words. A person's side is shown only where two separate readings of the quote agree; otherwise they are under "Side unclear". This page does not say who is right.
Agree 18 people
- ✓Theorem and Anthropic
Our experiments suggest that small AI models may be months away from being capable enough to self-replicate on commercial internet-of-things (IoT) devices
If small models improve at their historical rate, they’ll match this capability level by the end of this year.
A replicator like this would be challenging to shut down, since there’s no centralized point to shut off—though it may be slow (with a doubling time of days).
An AI botnet isn’t tied to any particular hardware or software type, as evidenced by small models being able to set up inference engines on machines from a Jetson Nano to a Mac
we expect a future awash with agents running on all types of hardware, including some that aren’t under human control.
- Quote not reprinted: this page already repeats as much of this article as we allow.
- Quote not reprinted: this page already repeats as much of this article as we allow.
lightweight agents may be much more willing than human attackers to put in the effort to compromise large fractions of the world’s IoT compute.
- ✓Jason Koebler
it has become clear that AI, broadly defined, has fundamentally changed both offensive hacking and defensive cybersecurity.
because AI is proving something that we have always known: much of the software that runs the internet, our devices, our accounts, and our lives is vulnerable, or out of date
AI and large language models have made it extremely easy to essentially brute force the internet
We’re about to find out how insecure the internet really is.
- ✓Exponential View
It will get worse before it gets better.
He is right.
they don’t analyse vulnerabilities the way humans do—they will look in places we might not.
Software, the internet, and cyberspace in general are currently under relentless attack.
- ✓Anthropic and OpenAI
we expect similar findings to increase sharply in the coming months
It highlights how AI can uncover security flaws with internet-scale impact, but the same capabilities can also help attackers rapidly develop offensive tools and exploits.
which may be in short supply if frontier capabilities are not too far ahead of those that are broadly available
- ✓Transformer
“About as well as it could have” is a long way from “well,” though.
But it probably does mean a very bumpy road ahead
We are probably in the best of all possible worlds in which AI has advanced cyberattack capabilities. The problem is that none of those worlds were ever good ones.
- ✓Sagi Tzadik
As the time and expertise required for this work decrease, complexity may no longer provide a meaningful barrier, changing how we define what is vulnerable in practice.
As AI reduces the effort needed to develop and adapt complex exploits, vulnerabilities that once seemed impractical to exploit become viable targets.
- ✓Latent Space
We should expect to be in for at least a year of finding such critical vulnerabilities
We should expect to be in for at least a year of finding such critical vulnerabilities;
- ✓Joshua Achiam
Joshua Achiam compares the moment to March 2020 , speculates about internet lockdowns, recommends backing up valued internet material, and suggests printing important records.
Joshua Achiam compares the moment to March 2020 , speculates about internet lockdowns, recommends backing up valued internet material, and suggests printing important records.
- ✓Ethan Mollick
Ethan Mollick interprets the moment as attack and defense accelerating beyond human response speeds
Ethan Mollick interprets the moment as attack and defense accelerating beyond human response speeds
- ✓Gary Marcus
they don’t need state-of-the-art “frontier” models to cause mayhem. This should serve as a wakeup call for those who focus all their attention on regulating “frontier models”
Things will get worse.
- ✓Kanishka Narayan
likely, if not inevitable,
- ✓Matthew Green
so many people discovering that the whole Internet is dry tinder and the only thing keeping it from burning is the fact that teenagers haven’t found the matches yet
- ✓Elias Taylor
- Quote not reprinted: this page already repeats as much of this article as we allow.
- ✓Jason Clinton
are going to get a lot worse before they get better.
- ✓Bill Gates
“tumultuous” 20 years of disruption from AI, calling it “the most dangerous weapon of all time”
- ✓Matthew Prince
“We’re going to have a couple of years where every single week it’s going to feel like the sky is falling.”
- ✓Jason D. Clinton
as bad as things have gotten in the past 4 days, they are going to get a lot worse before they get better.
- ✓Casey Handmer
I believe the computer security overhang is not a few dozen bugs that get patched and then we’re good, there are 100,000s of vulns that are extremely discoverable with infinite free tokens
Disagree 2 people
- ✗@BlancheMinerva
calls the "offense winning" narrative disinformation, arguing this was a defensive audit and there is no evidence of large-scale exploitation
- ✗Joshua Saxe
doesn’t immediately translate into damages.
No clear side 5 people
- ~Latent Space
it is the wrong move to blame the messenger for finding them; rather it is good to know that our critical software is generally getting the biggest security audit it has ever had in human history.
the defensible sequence is to obtain the authoritative PHP and distribution advisories, inventory exposed runtimes, apply temporary protections where justified, deploy authenticated fixed packages, verify running processes, and investigate evidence of exploitation separately.
it is the wrong move to blame the messenger for finding them; rather it is good to know that our critical software is generally getting the biggest security audit it has ever had in human history.
Unsolicited exploitation still risks unauthorized access, outages, incorrect fixes, destruction of evidence, and cross-jurisdictional liability.
“Black October” counts as such.
- ~Theorem and Anthropic
- Quote not reprinted: this page already repeats as much of this article as we allow.
- Quote not reprinted: this page already repeats as much of this article as we allow.
- Quote not reprinted: this page already repeats as much of this article as we allow.
- ~Zvi Mowshowitz
On a personal level, I would urge taking ordinary precautions in Peter Wildeford’s list.
My current read is that we are being far too conservative about who gets access to the tools for defenders.
- ~Anthropic and OpenAI
safeguards intended to prevent abuse can also block legitimate security work. Broad access to capable models and effective protections against misuse are both critical.
- ~Sagi Tzadik
The opportunity is to make this work routine: systematically finding flaws in existing software and catching new ones before they reach production.
Side unclear 9 people
- ·Alex Johnson
- Quote not reprinted: this page already repeats as much of this article as we allow.
- Quote not reprinted: this page already repeats as much of this article as we allow.
- ·J. D. Pressman
- Quote not reprinted: this page already repeats as much of this article as we allow.
- Quote not reprinted: this page already repeats as much of this article as we allow.
- ·Zvi Mowshowitz
The world seems like it is playing with fire. We have been saved, so far, almost entirely by the fact that People Don’t Do Things.
This last week of greater Hackening is plausibly in large part due to the new DeepSeek model.
- ·Import AI
these things are bound up, inseparably – the advance of AI systems in a general area like coding has also created both positive and negative impacts
- ·Jason Koebler
The answer in many cases is, probably, no.
- ·Elias Taylor
- Quote not reprinted: this page already repeats as much of this article as we allow.
- ·Dean W. Ball
- Quote not reprinted: this page already repeats as much of this article as we allow.
- ·Joshua Achiam
- Quote not reprinted: this page already repeats as much of this article as we allow.
- ·Kelsey Piper
- Quote not reprinted: this page already repeats as much of this article as we allow.
Week by week
- · 24: Import AI ·, Anthropic and OpenAI ✓, Sagi Tzadik ✓, Latent Space ✓, Joshua Achiam ✓, Ethan Mollick ✓, Alex Johnson ·, J. D. Pressman ·, Gary Marcus ✓, Jason Koebler ✓, Matthew Green ✓, Elias Taylor ·, Jason Clinton ✓, @BlancheMinerva ✗, Transformer ✓, Joshua Saxe ✗, Bill Gates ✓, Exponential View ✓, Matthew Prince ✓, Jason D. Clinton ✓, Zvi Mowshowitz ~, Casey Handmer ✓, Dean W. Ball ·, Kelsey Piper ·
- · 3: Theorem and Anthropic ✓, Zvi Mowshowitz ·, Kanishka Narayan ✓
Built 2026-11-21 06:50 UTC from the news knowledge
graph by code. Every quote was checked against the source's own words; which side a quote takes is a model's reading,
made twice.