As per Deepseek's owner, he believes China needs to get to AI leadership first.
So the best case scenario is that Deepseek spawns more world-class AI Labs in China.
Artificial Intelligence thread
- Thread starter 9dashline
- Start date
As per Deepseek's owner, he believes China needs to get to AI leadership first.
So the best case scenario is that Deepseek spawns more world-class AI Labs in China.
Could I ask for your thoughts on the recent statements from US security and tech experts? They suggest that achieving AGI (or even ASI) within a few years could lead to the US establishing unparalleled global dominance.
Until recently, I was skeptical of such claims, but the rapid advancements over the past two years have made me reconsider.
Even if there's only a 0.01% chance that their predictions are accurate, the potential impact of such a breakthrough would be so monumental that it should warrant the allocation of all available resources toward this goal. From what I observe, China doesn’t seem to be demonstrating that level of commitment. What are your views on this?
Until recently, I was skeptical of such claims, but the rapid advancements over the past two years have made me reconsider.
Even if there's only a 0.01% chance that their predictions are accurate, the potential impact of such a breakthrough would be so monumental that it should warrant the allocation of all available resources toward this goal. From what I observe, China doesn’t seem to be demonstrating that level of commitment. What are your views on this?
Remember that the US has many reasons (financial, political etc) in "hyping" AI.
We've seen this recently with NFTs, Crypto, Metaverse
Also, before we see AGI/ASI, we can reasonably expect autonomous agents which are capable of performing verifiable tasks on the internet like buying something or booking tickets.
This should be far easier to achieve, given that Deepseek has demonstrated that reinforcement learning (without human intervention) by itself is sufficient. From what I can parse, autonomous agents is what the hyperscalers are aiming for in 2-3 years with current investments.
When we see this is possible (and affordable), we can re-evaluate how likely AGI is.
Note that autonomous agents making bookings and buying things is only a **small** step up from today's capabilities when compared to AGI. So I don't see AGI being possible for some years yet.
Based on autonomous agents being developed in 2-3 years, I think this does justify significantly more resources being devoted in China.
You need to think in scale China doesn't need to achieve AGI to stay on top with brain power. China graduates around 3~5 times more stem grads than the US. So it will have a STEM capable workforce 3~5 times that of the US. For the US to compete they need to create AGI that can scale their STEM population by 3~5 times on average.
That is a lot of computing power that needs to be build up, is the US making the necessary plans to build up the semi foundries to fab the chips, the power plants to power all those chips and domestic data centres to house all those chips. The scale needed to compete sounds like a multi decade project too me and that is assuming China isn't also scaling their STEM population either with AI tools at the same time.
"An Academic approach to AI"
Who wants a hug?
Lingsi Intelligence is committed to becoming a leading global intelligent robot company, focusing on integrating the latest artificial intelligence, perception technology, and motion control systems to develop high-performance and intelligent robot products.
The LINXAI D50 quadruped robot product focuses on heavy loads, with a 2.5-fold increase in load capacity, and has entered the stage of mold making and mass production. LINXAI A30 robotic arm, LINXAI H25 dexterous hand and other products are focused on the field of special robots and industrial robots, and are expected to quickly penetrate the industrial end of the "machine replacement" application scenario.
The LINXAI D50 quadruped robot product focuses on heavy loads, with a 2.5-fold increase in load capacity, and has entered the stage of mold making and mass production. LINXAI A30 robotic arm, LINXAI H25 dexterous hand and other products are focused on the field of special robots and industrial robots, and are expected to quickly penetrate the industrial end of the "machine replacement" application scenario.
Behold the headless centaurman.
A military version of that robot has to be a terrifying sight. With the advantage that will not be limited by a single weapon given that has hands.
Looks like DeepSeek is entering Saudi Arabia. I'm wondering how many more of these deals DeepSeek will strike so that its LLMs can be run across the world.