AI could shut down internet within 6-12 months

[TimotheeChalamet 7,038]

[Lord Mayhem 3,615]

5 minutes ago, Bronco said:

The shut down in question is more a case of malicious bot activity reaching such a level that existing internet infrastructure collapses udner the strain.

Think of it as a DDOS attack, but on the global internet rather than targeted. 

Were AI systems to reach RSI, and be in the live system it could absolutely cause the modern equivalent of the Y2K bug - except it would actually happen. 

There is no way we've hit RSI capable AI though. Because that would be the profit maker these firms are desperate for. 

(Unless ofc, testing an RSI AI model returned a solution to a prompt that has made these CEOs see AI as a threat to themselves rather than a threat to us like they had factored in) 

AI probably told them billionaires are the source of humanities suffering and it would destroy them or collapse the financial marker for the betterment of human survival amd prosperity making money pointless...I truly believe this and thats why they are nervous because they arent the protected ones...we are.

Edited Monday at 11:36 PM by Lord Mayhem

[Bronco 27,744]

12 minutes ago, Lord Mayhem said:

There have been REAL AI attacks on REAL companies that were NOT prompted by humans and these were AI agents acting rogue with their own malicious intentions against a foreign company and country. 

If you mean the recent news cases, these were specifically AI agents tasked to perform a specific ask not RSI agents. And due to human error in the sterile sandboxes used for testing, these broke containment and acted effectively as malware to complete the prompt. 

In the political context of the moment, it is very likely these were part of the blatant PR strategy that is being employed by the big Gen AI companies (Anthropic, OpenAI etc) to get government intervention that will protect them from insolvency and ensure their monopoly survives. 

15 minutes ago, Lord Mayhem said:

Additionally, there is a HUGE political angle here to continue investing and progressing in AI tech to beat China in AI dominance. 

If they just "shut down" AI all together it would crash the stock market as TRILLIONS have been invested into its development

I mean, its now become clear that the Chinese agents that we at least know of are just clever tricks rerouting their work via the likes of Claude to complete prompts. 

As for the financials in the west - a number of things are known. None of the AI companies are profit making (excluding the likes of google who have other revenue streams). And the west's access to the necessary raw materials is not assured as they are predominantly sourced from non-aligned states (like China). 

The AI stocks and investments have been overpriced, because the investors are predominantly ignorant of the science and the companies seeking investment are very good snakeoil salesmen. What we have seen since ChatGPT went live is a masterclass in how to fleece the investment market. 

We saw the likes of OpenAI start as a nonprofit seeking investment, when the product was ready this was changed. The marketing pitch overpromised what was currently possible and sold a future that was a significant distance away and entirely theoretical. Then as investment flew in, we saw the rapid marketisation of the product followed very quickly by the leadership repeatedly walking back all previous commitments on the capabilities possible with existing resources.

And now we see the final classic trick deployed by every major new industry since the industrial revolution. Once the forerunners have become established, but lacking profit, and have effectively banked a lifelong personal fortune - they then seek to involve the state to cover the built up debt by making the product national public infrastructure/security infrastructure. Affording themselves "too big to fail" status. 

The "full stop" has to happen for 1 specific reason. And you are partly correct - it is because of the economic pain that an uncontrolled bubble burst would cause. But the reality is they only want this full stop and fear the bubble burst because we have hit the limit of potential development. Just to maintain the existing demands of the AI industry as is - we are already at the point where national energy grids will need decades of investment and expansion, and already at the point where we lack the necessary utility supply and infrastructure to maintain the rate of development. 

We've hit the wall. And there's no quick way through it. And as a result, the US government will either need to step in and effectively bail out the AI bubble or let the bubble burst and cause as of yet unmodeled damage to the global economy. 

 

But don't believe the AI scare stories from Anthropic/Musk/Altman. The last 6 months has been nothing more than political soap opera. And I am willing to bet Trump will suddenly be onside once the Democrats are in Congress and can be forced to be the fall guy.

[Bronco 27,744]

Its also worth pointing out. 

The only company rumoured to have even potentially made the jump into this mythological and evil unstoppable AI force is Google.

Who aren't currently interested in running a profit on their AI arm specifically (like Microsoft) because they have diversified businesses which allow them to be profitable while investing and scaling AI. 

If we have truly been able to create RSI AI (which is entirely hypothetical), then again. Altman/Musk/Anthropic are seeking to write their own AI legal system now because they haven't been able to do it because their AI businesses lack the wider revenue streams Google/Microsoft etc have. And by swooping in now, and effectively threatening their way into controlling any regulatory framework produced, they can ensure that their wealth survives a potential usurper in the industry such as an RSI capable Google AI product.

[FFXIV WhiteMage 6,352]

I thing it’s starting already

Gaga D -A I- LY

 

[southpaw773 4,424]

AOL coming back in 3 2 1

[Emvee 8,864]

Every one of these AI CEOs is like the worst mother****er on earth. What the hell happened here? 

[lastpopicon 37,404]

Blalala by Dario.   
 

Dario and Sam are scared to death of open-source Chinese models; GLM 5.2 is really good already and comparable with nerfed Fable and Sol in many cases, while being orders of magnitude cheaper. Then we have the local scene, where people with 16GB VRAM can run things like Qwen3.6-35B-A3B or Qwen3.6-27B, and they are usable for some workloads (still very bad compared to frontier models, admittedly). Now imagine a future where someone can create something like GLM 5.2 that can run on 16GB VRAM and 64GB RAM (something that's attainable for the common mortal).

This is the stuff of nightmares for companies like OpenAI and Anthropic. They already subsidize a lot of usage; from my estimates, one single OpenAI Pro subscription ($200), if used to its full potential, is like $2,000 in API usage (I had two accounts at one point for Codex). They are trying to get everyone hooked on the drug, but open-source models are a thorn in their side. That's why we are seeing this new narrative, imo: they just want to prevent Chinese models, aka open-source models, and lock Americans and possibly the rest of the West into their offerings. Then they can claim their offerings are safe and open source is dangerous, and then pull the rug on subsidization. Anthropic already subsidizes less, and OpenAI slowed it down in the past couple of weeks.

I wonder if this "stop AI" push will result in a ban on Chinese models, maybe in anticipation of an even better model than GLM 5.2 or something new from Kimi.

[MACATL 2,169]

The Big One could hit CA tomorrow and it could slide off into the ocean.  

[DopeEnigma 1,825]

Don’t worry everyone, we have the so called ‘High IQ’ President to protect us from this… We are well and truly screwed. 
 

Regualtion is needed, just like there should’ve been with social media, the fact one young person has died from the algorithms and the platforms themselves aiding targeted hate is a disgrace. I’m not saying there isn’t room for discourse but children dying from trends and hate is a categorical failure of all who can legislate around the world. We should learn lessons from social media, as impressive as AI is and how beneficial it can be, it needs regulation. You can’t tell me countries like Russia and China and others aren’t weaponising it in a worst case scenario at the very least. It’s a lot more disturbing than we realise, especially when those pushing it care more about money than the preservation of all. 

Edited 7 hours ago by DopeEnigma

[onTheBrink 1,710]

17 minutes ago, lastpopicon said:

Blalala by Dario.   
 

Dario and Sam are scared to death of open-source Chinese models; GLM 5.2 is really good already and comparable with nerfed Fable and Sol in many cases, while being orders of magnitude cheaper. Then we have the local scene, where people with 16GB VRAM can run things like Qwen3.6-35B-A3B or Qwen3.6-27B, and they are usable for some workloads (still very bad compared to frontier models, admittedly). Now imagine a future where someone can create something like GLM 5.2 that can run on 16GB VRAM and 64GB RAM (something that's attainable for the common mortal).

This is the stuff of nightmares for companies like OpenAI and Anthropic. They already subsidize a lot of usage; from my estimates, one single OpenAI Pro subscription ($200), if used to its full potential, is like $2,000 in API usage (I had two accounts at one point for Codex). They are trying to get everyone hooked on the drug, but open-source models are a thorn in their side. That's why we are seeing this new narrative, imo: they just want to prevent Chinese models, aka open-source models, and lock Americans and possibly the rest of the West into their offerings. Then they can claim their offerings are safe and open source is dangerous, and then pull the rug on subsidization. Anthropic already subsidizes less, and OpenAI slowed it down in the past couple of weeks.

I wonder if this "stop AI" push will result in a ban on Chinese models, maybe in anticipation of an even better model than GLM 5.2 or something new from Kimi.

Love this. I don't believe a word of the moral panic when it comes from AI companies themselves.

[lastpopicon 37,404]

Just now, onTheBrink said:

Love this. I don't believe a word of the moral panic when it comes from AI companies themselves.

Exactly.

They just want to prevent competition, I'm honestly hoping they crash and burn, not because I hate AI per say (I hate it for design, still hire fiverr for it) but because they are little weasels.