5 questions for Christopher DiCarlo
Derek Robertson, Politico
What’s one underrated big idea?
I’ve spent my life studying and teaching and researching critical thinking, and I think the implementation of critical thinking into our educational systems needs to be ramped up.
When we abandon values of responsible human reasoning, we do so at our own peril. Once logic and facts, and the foundational aspects of critical thinking lose favor or proper influence over human communication, any working system of government is destined to implode. We need to introduce a critical thinking skill set at all levels of education that would be packed with tools that allow people to diagram and construct arguments intelligently, to recognize and account for their biases and the biases of others, to appreciate context, to establish appropriate and relevant evidence, and to be able to spot fallacies or errors in reasoning.
The more widespread literacy and critical thinking are, the less the effect will be from the malicious spreading of misinformation, disinformation and conspiracy theories which warp humanity’s views away from rationality and more towards emotionally charged groupthink.
Is Alibaba’s Qwen the Open-Source AI Winner?
Michael Spencer and Grace Shao
The Latest Qwen Model Surpasses some Proprietary Closed-Source Models
On November 11th, 2024 Alibaba Cloud’s Qwen team released Qwen2.5-Coder-32B-Instruct. The performance of this model surprised a lot of people but not those of us who were already bullish on Qwen. The flagship model, Qwen2.5-Coder-32B-Instruct, reaches top-tier performance, highly competitive (or even surpassing) proprietary models like GPT-4o, in a series of benchmark evaluation, including HumanEval, MBPP, LiveCodeBench, BigCodeBench, McEval, Aider, etc. It reaches 92.7 in HumanEval, 90.2 in MBPP, 31.4 in LiveCodeBench, 73.7 in Aider, 85.1 in Spider, and 68.9 in CodeArena.
Impact Assessments are the Wrong Way to Regulate Frontier AI
Dean W. Ball, Hyperdimensional
But this analysis also points the way toward the easiest way to improve these bills. Specifically, the low-hanging fruit for improvement is:
- Completely exempt all generative AI foundation models (like ChatGPT) from the law. Some versions of this law already have gestures at this exemption, but none I have seen achieve the full exemption that is appropriate.
- Change the definition of “algorithmic discrimination” to be focused on discriminatory intent rather than disparate impact. This will narrow the focus of the law.
