nsuchaud – insights that matters

State of AI 2023 – Report

  1. Research: Technology breakthroughs and their capabilities.
  2. Industry: Areas of commercial application for AI and its business impact.
  3. Politics: Regulation of AI, its economic implications and the evolving geopolitics of AI.
  4. Safety: Identifying and mitigating catastrophic risks that highly-capable future AI systems could pose to us.
  5. Predictions: What we believe will happen and a performance review to keep us honest.
  1. GPT-4 is the master of all it surveys (for now), beating every other LLM on both classic benchmarks and exams designed to evaluate humans, validating the power of proprietary architectures and reinforcement learning from human feedback.
  2. Efforts are growing to try to clone or surpass proprietary performance, through smaller models, better datasets, and longer context. These could gain new urgency, amid concerns that human-generated data may only be able to sustain AI scaling trends for a few more years.
  3. LLMs and diffusion models continue to drive real-world breakthroughs, especially in the life sciences, with meaningful steps forward in both molecular biology and drug discovery.
  4. Compute is the new oil, with NVIDIA printing record earnings and startups wielding their GPUs as a competitive edge. As the US tightens its restrictions on trade restrictions on China and mobilizes its allies in the chip wars, NVIDIA, Intel, and AMD have started to sell export-control proof chips at scale.
  5. GenAI saves the VC world, as amid a slump in tech valuations, AI startups focused on generative AI applications (including video, text, and coding), raised over $18 billion from VC and corporate investors.
  6. The safety debate has exploded into the mainstream, prompting action from governments and regulators around the world. However, this flurry of activity conceals profound divisions within the AI community and a lack of concrete progress towards global governance, as governments around the world pursue conflicting approaches.
  7. Challenges mount in evaluating state of the art models, as standard LLMs often struggle with robustness. Considering the stakes, as “vibes-based” approach isn’t good enough.