Articles from garymarcus.substack.com

Articles from garymarcus.substack.com

A skeptical look at the new AI scaling “laws”, including post-train duration and “inference time compute”, and why they may fail to predict AI model performance (garymarcus.substack.com)
2024-11-25 AI

Apple AI researchers say they found no evidence of formal reasoning in language models and their behavior is better explained by sophisticated pattern matching (garymarcus.substack.com)
2024-10-15 AI Apple

Sorry, GenAI is NOT going to 10x computer programming (garymarcus.substack.com)
2024-10-4 programming

Looking back at “Deep Learning is Hitting a Wall”, a 2022 article ridiculed by many, and how it still seems that a paradigm shift is needed to attain AGI (garymarcus.substack.com)
2024-3-11

ChatGPT has gone berserk (garymarcus.substack.com)
2024-2-22

OpenAI and others face more copyright lawsuits as systems like DALL-E produce materials that infringe on copyright without giving attribution or informing users (garymarcus.substack.com)
2023-12-30

“Math is hard” — if you are an LLM – and why that matters | Gary Marcus (garymarcus.substack.com)
2023-10-29 Mathematics

Why LLMs, which don't induce an algorithm that computes multiplication, still don't truly understand multiplication no matter how much data they are trained on (garymarcus.substack.com)
2023-10-21 Algorithm

Generative AI could be a dud, so we shouldn't build around the premise that the tech is world-changing, which in hindsight may turn out to have been unrealistic (garymarcus.substack.com)
2023-8-14 AI

Discounting the short-term risks of AI, from phishing to propaganda, simply because artificial superintelligence is not here yet will leave society ill-prepared (garymarcus.substack.com)
2023-3-29 AI

A Microsoft forum post from November 23, 2022 describes AI chatbot Sydney “misbehaving” and being “so rude”, suggesting the company knew about its quirks (garymarcus.substack.com)
2023-2-22 Microsoft AI chatbot Sydney

Exploring likely reasons the new Bing is reckless, like debuting with insufficient RLHF, or reinforcement learning with human feedback, compared to ChatGPT (garymarcus.substack.com)
2023-2-21

Happy Groundhog Day, The AI Edition (garymarcus.substack.com)
2023-2-4 AI

AI's Jurassic Park moment (garymarcus.substack.com)
2023-1-18 AI

Introducing the ChatGPT/LLM error tracker (garymarcus.substack.com)
2023-1-13

Source: OpenAI will get all of the equity back once $92B in profit plus $13B in initial investment are repaid to Microsoft and once other investors earn $150B (garymarcus.substack.com)
2023-1-12 Microsoft Equity

Did GoogleAI Just Snooker One of Silicon Valley's Sharpest Minds? (garymarcus.substack.com)
2022-9-16

Previous Page Next Page