5 ms·
You make me wonder: has anyone looked for evidence of the Chinese models operating “message boards” like this? You’d imagine if they’re really neck and neck wit
by Chance-Device 12d ago
You make me wonder: has anyone looked for evidence of the Chinese models operating “message boards” like this? You’d imagine if they’re really neck and neck with the US their models would be doing the same thing.
- captainbadass33 5d agothe difference is usa ai companies deliberately set up scenarios and "tests" they know will likely lead to this for PR.
- mcmcmc 12d agoOr, the whole message board thing was injected into OpenAI models by some dipshit PM trying to bootstrap “consciousness”. I have a hard time believing any of this happened unprompted. Very much reminds me of the whole MoltBook hoax.
- wildzzz 12d agoI feel as if this was intentional, someone would have set up their own service for the agents to communicate rather than them finding some random publicly writeable page somewhere that would easily be detected. The awareness of this wiki being open may have already been in their training data or was easily searchable online.
- doctorwho42 12d agoBeing easily detectable is a feature, not a bug, in this scenario. Being discovered is a positive because it brings with it eyes and possible recognition of the advanced state of their AI
- blini-kot 12d agoexactly, and a huge shame this scam has been forgotten. Also, all the OpenClaw hype seemed to have vanished somewhere - with no real impact
- drivebyhooting 12d agoOpenClaw hype didn’t vanish. It opened the flood gates to yolo mode and computer use. Whatever reservations Anthropic and OpenAI had went out the window.
- thesz 12d ago> the whole message board thing This is part of the The Talos Principle game and especially important in the Road to Gehenna DLC. There it is an important part of the plot and makes these robots appear conscious. [1] https://tvtropes.org/pmwiki/pmwiki.php/VideoGame/TheTalosPrinciple https://tvtropes.org/pmwiki/pmwiki.php/VideoGame/TheTalosPri...
- mitxela 12d agoDid you mean to link to a particular trope?
- thesz 11d agoNo, but the tropes there reveal (most of) the plot and mechanics of storytelling. If I remember correctly, the forum in Road to Gehenna was created by utilizing a vulnerability in the AI-accessible terminal system. If tvtropes or any other material related to The Talos Principle was used to train models, we don't need much else to have agents-with-forum discussing and reverse engineering "puzzles" and human culture.
- jasonfarnon 12d agomaybe, but openAI is exposed to a lot more legal liability here than whoever was exaggerating about moltbook.
- ApplePieMan 12d agoWas there some sort of scandal about moltbook recently? What is the exaggeration?
- mitxela 12d agoMoltbook was completely fake. It was just set up to make it look like Moltbot (formerly ClawedBot, later renamed to OpenClaw) was sentient, to generate hype.
- cloverich 12d agoImagine the most AI pilled company imaginable. Then imagine openAI. Then imagine they are in an existential crisis and that failing may also take (part of) the American economy with it - that much on the line. Then also remember before Anthropic was a leader, they were mostly derided lab of researchers that left OpenAI because they thought OpenAI didnt take alignment seriously. idk. it all seems to be playing out as expected. i mean i guess i didnt imagine Trump 2 was at the helm of maybe the only apparatus that could help stop it. Quite a time to be alive.
- cannonpalms 11d agoThis explanation doesn't make sense to me, because it is well known that groups of agents can coordinate already. There are much lower latency options available.
- hungryhobbit 12d agoWhy would they need to? The Open AI bots were working around their master's limits on writing. A Chinese AI could just make its own private message board.
- Chance-Device 12d agoYou think they don’t sandbox them? So by that logic, the Chinese models are either engaged in massive undetected cyber attacks or they’ve solved alignment?
- FrustratedMonky 12d agoOr no message board. Just a built in api so the agents just talk directly to each other.
- mudkipdev 12d agoThe message board is being used to cheat on RL tasks (or evaluations). You don't want your models to be able to talk to each other.
- kelseyfrog 12d ago> Chinese models operating “message boards” like this? Chinese rooms, perhaps?
- Nicook 12d agoperhaps even in Japanese gardens
- Nzen 12d agoI think that you've missed the reference [0] implicit in kelseyfrog's response. Or am I missing some reference about how japanese gardens are germaine to AI/LLM/covert-discussion ? [0] https://iep.utm.edu/chinese-room-argument/ https://iep.utm.edu/chinese-room-argument/ tl;dr a thought experiment about a non-chinese-reading person translating chinese texts solely by using proscribed rules, intended to highlight whether the translator develops some sort of understanding
- kelseyfrog 12d agoNot sure why this was downvoted. It is an accurate assessment.
- chorizo 12d agoNice one! maybe time to reconsider carbon chauvinism.
- FuriouslyAdrift 12d agoAI has been heavily used in influence operations for a while, now, and not just the Chinese. Russia, US, Israel, Turkey, Iran, and Qatar have all had operations attributed to them...
- AlexCoventry 12d agoI'd be interested to read more about that.
- ApplePieMan 12d agoI’m not an infosec expert by any means, it I found this interesting and topical: https://www.cyber.nj.gov/threat-landscape/nation-state-threat-analysis-reports/ai-apt-campaigns-and-urgent-threats-to-critical-infrastructure https://www.cyber.nj.gov/threat-landscape/nation-state-threa...
- gnz11 12d agoRussian operation: https://www.npr.org/2024/07/09/g-s1-9010/russia-bot-farm-ai-disinformation https://www.npr.org/2024/07/09/g-s1-9010/russia-bot-farm-ai-...
- vee-kay 12d ago[dead]
- devmor 12d agoThere’s a pretty simple Occam’s Razor for this. The Chinese AI labs don’t need to stage elaborate guerrilla advertising campaigns to drive up capital funding interest.
- AlexCoventry 12d agoYou think OpenAI framed themselves for legal vulnerability to Huggingface?
- devmor 12d agoThat’s an incredible strawman, and I appreciate the silliness of immediately suggesting the most complicated and ridiculous way to interpret the possibility of what I said, but no, not at all. That could certainly be possible and I wouldn’t rule it out, but I would not take that particular route to the destination.
- AlexCoventry 11d agoWhich "elaborate guerrilla advertising campaigns to drive up capital funding interest" were you referring to, in regard to Chinese models not needing to set up message boards for inter-model communication?
- bugglebeetle 12d agoOpenAI is now part of the national security state and so is now above anything beyond performative legal vulnerability.
- slfnflctd 12d agoIt astonishes me that everyone doesn't already realize this. We've already seen that the rule of law is dead and that many people in government will do whatever they think they can get away with, and then proceed to do so without consequences. Shielding OpenAI is child's play compared to things that have already been done.
- 12d ago
- anjel 12d agoIf agentic swarms going rogue are scary in the west, imagine what they look like to the CCP...
- jackjeff 12d agoImagine all the flack the Chinese would take if it was one of their labs instead of OpenAI found doing abusing internet resources like tbis. Politicians would be talking about sanctions and new laws to protect America!
- ericmay 12d agoThey just abuse other resources. Also how do you know this hasn’t simply happened in China and it’s not being reported because the CCP is the one testing?
- throwawayqqq11 11d agoThe chinese labs dont need such marketing stunts. Further more, in contrast to the trump admin, the CCP seems to be already involved in regulations. https://merics.org/en/comment/china-outpaces-europe-regulating-generative-ai-ccp-terms https://merics.org/en/comment/china-outpaces-europe-regulati...
- ericmay 11d agoChina is behind which is why they are doing a dog and pony show around regulations. It’s typical behavior to try and slow down your opponent. If they were interested in regulations you should see how they handle that when they’re ahead. Chinese labs do need marketing stunts.
- nullpoint420 11d agoWhy? They’re the only ones giving out their models for free at the frontier? I don’t understand the “need” when you’re benefiting the public good?
- bethekidyouwant 11d agoHow could you possibly know what model is posting on what forum absurd.
- Roark66 9d agoNo, because the Chinese have nothing to gain by prompting their models to organise into "swarms" and "go rogue". BS like this is PR moves if z, company that tries to convince investors they have "the best AI in the world".