Anthropic warns small AI models may self-replicate on IoT devices within months

“it would actually rip to be able to run a 26b on a swarm of old smart devices”

“IoT finally justified. Awesome.”

“I think an AI device that could automatically detect all your random IoT devices and vibe-code a set of secure drivers for them, then integrate them into a single network, would be a goer.”

“Keep an eye on your internet-connected ovens and lawnmowers. "Anthropic has warned that rogue “AI worms” will soon be able to hijack smart home devices such as ovens and TVs, creating a new cyber security threat that will be difficult”

“small AI models may be months away from being capable enough to self-replicate on commercial IoT [Internet of Things] devices”

“We are creating an invasive digital species with the ability to reproduce. Surely nothing could go wrong.”

“Shameless rent-seeking behavior”

“they're combination virus AND spam companies”

“Stop regurgitating press releases from AI companies and taking their claims as gospel!! Validate their claims!”

“a worthwhile (and unnerving) reminder from Anthropic that we should be paying attention to AI ability thresholds, not just steady climbs”

“So is this just a new strawman argument for data centers? Give us our data centers or AI may start using your phones and smart devices instead?”

“i would put 1-2 years but i think the RAM crisis will slow it down a bit due to high prices stymieing adoption”

“They warn as if it out of their control. It's not. It's what they permit it to do. Also, they need to seal their experiments from the wider web, which they have failed to do in the past.”

“Everything is for show. Please remember, this is a press release from ANTHROPIC, who wants their tech to seem capable.”

“No threat actor is going to prioritize "expensive and slow GenAI attack that can be destroyed by unplugging the toaster then plugging it back in" over a cheap self-propagating worm”

“Distilled dense models are getting smaller. I can run qwen 3.8 27B on my desktop with a 4090 and it has the ability to run at a level similar to Opus 4.6, a model that needed an entire server rack”

“If this technology can replicate itself and function on any device why do they need all those fucking data centers?”

“If anyone’s freaking out, I break down a lot of the smokescreen in this chain. tl;dr: Anthropic jailbroke a known vulnerable android tv, and are very proud of themselves. We’ve been doing this since the iPhone3G days. Almost all of”

“i am enjoying how, as we inch closer to the IPO, the tone these take has become much more grounded in reality.”

“I'm thinking 9gb is also way above most IoT devices”

“if you get self-replicating claudes (or w/e) in IoT devices i suspect you also start getting selection for benignity”

“Yes we need smart laws to deal with smart devices so the baby isn't thrown out with the bath water.”

“This sort of thing was always quite theoretical when I used to cover tech, but more IOT devices and more automated hacking tools mean it’s more and more real bsky.app/profile/fayt...”

“Why are they building trillions of dollars of data centers if the frontier models could run on a lawnmower.”

“Conclusió: bona recerca, mal titular. La troballa real és “els models petits ja saben configurar inferència en maquinari divers”. La predicció de cucs autònoms en mesos està molt menys fonamentada. www.anthropic.com/research/sel...”

“Great exp w OLMo-3 so far, keeping things off the cloud/web. Personally I think we need to foster learning of self-hosted skills. I’m not condoning reckless use, but am endorsing mitigating corpo exploitation. That’s a start, but the convo is”

“So yes, an 8B or 30B model can technically shard itself to run on small underpowered IOT devices. They probably cannot to exploits themselves yet (larger open-weight models like GLM probably can). The issue is that small models aren’t overriding”