Source URL: https://simonwillison.net/2025/Mar/19/my-thoughts-on-the-future-of-ai/
Source: Simon Willison’s Weblog
Title: My Thoughts on the Future of "AI"
Feedly Summary: My Thoughts on the Future of “AI"
Nicholas Carlini, previously deeply skeptical about the utility of LLMs, discusses at length his thoughts on where the technology might go.
He presents compelling, detailed arguments for both ends of the spectrum – his key message is that it’s best to maintain very wide error bars for what might happen next:
I wouldn’t be surprised if, in three to five years, language models are capable of performing most (all?) cognitive economically-useful tasks beyond the level of human experts. And I also wouldn’t be surprised if, in five years, the best models we have are better than the ones we have today, but only in “normal” ways where costs continue to decrease considerably and capabilities continue to get better but there’s no fundamental paradigm shift that upends the world order. To deny the potential for either of these possibilities seems to me to be a mistake.
If LLMs do hit a wall, it’s not at all clear what that wall might be:
I still believe there is something fundamental that will get in the way of our ability to build LLMs that grow exponentially in capability. But I will freely admit to you now that I have no earthly idea what that limitation will be. I have no evidence that this line exists, other than to make some form of vague argument that when you try and scale something across many orders of magnitude, you’ll probably run into problems you didn’t see coming.
There’s lots of great stuff in here. I particularly liked this explanation of how you get R1:
You take DeepSeek v3, and ask it to solve a bunch of hard problems, and when it gets the answers right, you train it to do more of that and less of whatever it did when it got the answers wrong. The idea here is actually really simple, and it works surprisingly well.
Tags: generative-ai, deepseek, nicholas-carlini, ai, llms
AI Summary and Description: Yes
Summary: The text presents Nicholas Carlini’s insights on the future of AI, particularly regarding language models. It explores the potential advancements and limitations of LLMs and emphasizes the importance of keeping an open mind about their capabilities.
Detailed Description:
Nicholas Carlini, initially skeptical about the efficacy of large language models (LLMs), reflects on their future and proposes that significant advancements are likely in the coming years. His analysis involves an examination of the potential directions that AI technology could take, along with associated implications for the field.
– **Key Messages:**
– **Diverse Futures**: Carlini suggests that LLMs might either achieve capabilities surpassing human experts in economically useful cognitive tasks or evolve incrementally without a transformative leap.
– **Error Bars on Predictions**: He emphasizes the uncertainty in forecasting where AI technology will head, asserting the importance of maintaining wide error bars to account for various potential outcomes.
– **Concept of Limitations**: Carlini holds that there will likely be fundamental limits to the capabilities of LLMs, although he admits he cannot currently identify what these limits may be.
– **Example of Improvement**: He illustrates a method for enhancing LLM performance using a model called DeepSeek v3, which involves solving harder problems and adjusting training based on performance to improve task handling.
– **Implications for Security and Compliance Professionals**:
– **Vigilance in Progress**: As AI technology evolves unpredictably, security and compliance professionals need to stay updated on advancements to preemptively address potential vulnerabilities.
– **Understanding Limitations**: Recognizing the potential limitations of LLMs can guide the development of security measures that are robust against unforeseen challenges.
– **Integration into Practices**: The insights regarding the evolutionary path of LLMs should inform practical implementations and governance strategies related to AI security, helping organizations adopt resilient infrastructures that incorporate emerging technologies while mitigating risks.
This analysis underscores the complexity of AI’s evolution and its implications for security and compliance in technology deployment.