This is exactly my view and it's frustrating that a sizeable group of people refuse to accept there's any risk with AI.
We're literally trying to build something that's smarter than all humans and with the capacity to do extraordinary work while most people think there's no risk in open sourcing this capability with no guardrails (model ablation means no guardrails and as far as I'm aware can't be stopped).
Has our distrust in western companies really gotten to the point they'd rather unfiltered Astra level capabilities in the hands of every hostile government and terrorist organisation than potentially do any form of regulation in case it plays into the AI labs hands.
What's an example of him crying wolf? Most of the things have either been longer term worries or short term things that have come true (eg, the worries over gpt2 were about spam and possible impersonation, which came true. His prediction that LLMs would write most code and humans would still guide them has also come true in most orgs)
> Has our distrust in western companies really gotten to the point they'd rather unfiltered Astra level capabilities in the hands of every hostile government and terrorist organisation than potentially do any form of regulation in case it plays into the AI labs hands.
Ever checked political programs of Thiel, Musk and the rest of billionaires? They are literally and openly hostile to democracy, human rights, freedom and govermantal structures that made world more peaceful.
I'm tired of this rationale. "I'm more worried about grandma than $CORPORATION'S monopoly abuses" is exactly how you get manipulated by these billionaires. They want you to be addicted to their power.
Uncensored Chinese LLMs will be available to scammers forever now. They're open-weight and finetunable. You don't get to control that, and the self-serving billionaire assholes aren't capable of changing it either. They're only looking out for themselves.
I don't think Pacing is talking about current models. It's talking about future models, if development continues, that may theoretically be smarter than everyone
It's also a boring retort. Current LLMs can be faster and likely smarter than any one person when it comes to hacking. We're already in the endgame, as far as real-world risks go.
Why does everything have to be so black and white? It always either there's no threat at all or we all need to panic. How about be open to a reasonable discussion on potential outcomes and ways we can minimise risk?
There's some irony here because despite how many times climate change has been mentioned in this thread the current reaction mirrors climate change discourse with the majority of the thread denying the possibility of real AI risks and not even considering it as an intellectual question.
Hours spent at work have also gone down over the decades and leisure time has gone up. Even if that stat wasn't true and we all spent the same amount of time working more productively then that would still be good for society at large.
This attitude of unwillingness to do anything except for direct personal gain isn't good for the world and isn't good for yourself.
Hours spent working haven't noticeably declined for people in Northern Europe for fifty years. I worked 40 hours a week in the UK in 1978, the usual working week in Norway now is 37.5 hours.
I very much doubt that most people are greatly more productive, if you can even measure that for most jobs, than they were fifty years ago. Most jobs now are office jobs and service jobs. Much of the purported gains in productivity are soaked up in increased communication effort and in shifting ever more tat.
> I very much doubt that most people are greatly more productive, if you can even measure that for most jobs, than they were fifty years ago. Most jobs now are office jobs and service jobs. Much of the purported gains in productivity are soaked up in increased communication effort and in shifting ever more tat.
The old jobs are a lot more productive. We just decided to consume a lot more, and also a lot more of things that are harder to make more productive, instead of taking extra leisure.
If you were willing to live a 1950s life (and it were legal), you could mostly do that on very little money today.
That's how the theory goes but I would argue the observable evidence says otherwise. TSMC is Taiwan's national champion and effectively a global monopoly, Samsung is the same for Korea and is highly effective on the world stage, same again with ASML.
When a country does what it can to support its key industries those companies often out compete their global rivals and this is exactly the strategy China is using against the west with their support of domestic electric car manufacturing, solar panels, battery tech.
Alternatively the west is often anti big business and we've repeatedly seen national championions go under.
The savings in CO2 emissions for the likes of rapeseed, sunflower etc per MJ of energy is only around 50% compared to diesel. You have to release a lot of CO2 to grow these crops intensively. And that’s before you even consider where the land for these crops come from, potentially chopping down CO2 absorbing rainforests in many sorts of the world.
Biofuel from Used Cooking Oil is worthy considering for sure, but the quantities possible are tiny compared to the petrol/diesel we use.
The original iPhone cost $500 and adjusting for inflation that is about $800 today. I don't remember phones costing $300 since before the smart phone era. Even Walkman phones cost $500 in the early 2000's.
You do know the modern smartphone is the most technologically advanced consumer good that has ever existed? Like at no point in history has there ever been something so small, complex and intricate in the hands of so many people. The new iPhone is going to be very expensive - it’s a luxury item and marketed like that.
They can definitely make phones for $300 and you can buy a smartphone that is more powerful than the original iPhone for $50. It probably will be a bit slow running modern apps but it may possibly still be workable for a while.
It fundamentally has the same capabilities as the original iphone. You can browse the web, check your emails, watch movies, play simple games.
The only difference is that now you can play graphicially intense desktop-style games, but no one really does that. And also it has some sattelite phone functionality. The cameras are also better, the battery life is better, but it all doesnt fundamentally change the way you use it. The complexity over last year's smartphone is unnessisary.
Now that everyone in the world is basicially required to use a smartphone, the only way for big smartphone to make more revenue is to sell each consumer more smartphone per smartphone.
You're correct. Ultimately 99% of the stuff you do on a phone is identical to 10 years ago. I know exactly 0 zero phone gamers. I do know some who play, like, the pokemon trading card game.
Cameras have improved, but less than you think. Most of the improvement is in low-light, but in well lit scenes we actually sort of peaked 10 years ago. It's gotten slightly worse now, believe it nor not.
But you can't legitimately run a phone from 10 years ago, because while everything is the same, the software uses 100x as many resources. That sort of tricks people into thinking "oh it must be doing so much more!"
But... is it? Think back. Instagram 10 years ago, YouTube 10 years ago. Yeah... it's basically the same. Some argue they've regressed. I'm sure we've more than 10x the code size, but nobody sees that.
Web browsers have also gotten much more complex, but again, you won't notice it. We have a lot more fringe web APIs like webusb but like... the regular web browsing experience for 99% of tasks is identical.
I think there have been some niceities in web apis like webgpu, websocket, webrtc, PWAs. But they are limited in weird ways, and the fingerprinting has never been worse
Sorry but your whole argument that Anthropic must be burning cash is extremely weak and falls into the same tropes as Zitron.
We can actually get a very good approximation of the gross margin Anthropic sells their models over API by estimating model sizes which is pretty easy with comparible open source models and analysis points to 80-90% margins.
It also make sense for Anthropic to be profitable whilst raising large amounts of capital. This is because we're in a regime where just deploying GPU compute provides a huge upside and also demand for that compute has been relentless. Given that Anthropics best move would be to expand as fast as possible which requires raising capital as otherwise their organic growth would be far to slow.
There's about 20 million software engineers in the west and almost 50 million globally. Now add to the equation non technical paying subscribers who migrated from ChatGPT and enterprise API spend from internal AI applications and the figure seems pretty reasonable.
reply