ByteDance officially launches its latest Doubao large model 1.5 Pro (Doubao-1.5-pro), which demonstrates outstanding comprehensive capabilities in various fields, successfully surpassing the well-known GPT-4o and Claude3.5Sonnet in the industry. The release of this model marks an important step forward for ByteDance in the field of artificial intelligence. Doubao 1.5 Pro adopts a novel sparse MoE (Mixture of Experts) architecture, utilizing a smaller set of activation parameters for pre-training. This design's innovation...
While I don’t know whether this applies to DeepSeek R1, the Internet perpetuates many human biases and machine learning will approximate and pick up on those biases regardless of which country is doing the training. Sure you can try to tell LLMs trained on the Internet not to do that — we’ve at least become better at that than Tay in 2016, but that probably still goes about as well as telling a human not to at best.
I personally don’t buy the argument that you should hate the designer instead of the technology, in the same way we shouldn’t excuse a member of Congress’ actions because of the military-industrial complex, or capitalism, or systemic racism, and so on that ensured they’re in such a position.
I don’t see these tools replacing humans in the decision making process, rather they’re going to be used to automate a lot of tedious work with the human making high level decisions.
The fact that there is nuance does not preclude that artifacts can be political, whether intentional or not..
While I don’t know whether this applies to DeepSeek R1, the Internet perpetuates many human biases and machine learning will approximate and pick up on those biases regardless of which country is doing the training. Sure you can try to tell LLMs trained on the Internet not to do that — we’ve at least become better at that than Tay in 2016, but that probably still goes about as well as telling a human not to at best.
I personally don’t buy the argument that you should hate the designer instead of the technology, in the same way we shouldn’t excuse a member of Congress’ actions because of the military-industrial complex, or capitalism, or systemic racism, and so on that ensured they’re in such a position.
I don’t see these tools replacing humans in the decision making process, rather they’re going to be used to automate a lot of tedious work with the human making high level decisions.
That’s fair, but human oversight doesn’t mean they’ll necessarily catch biases in its output
We already have that problem with humans as well though.
There’s value in the tedious decisions though
The tedious decisions are what build confidence and experience
People build confidence doing work in any domain. Working with artificial agents is simply going to build different kinds of skills.