Hard not to be awed reading this. Teaser: “AGI by 2027 is strikingly plausible. GPT-2 to GPT-4 took us from ~preschooler to ~smart high-schooler abilities in 4 years. Tracing trendlines in compute (~0.5 orders of magnitude or OOMs/year), algorithmic efficiencies […]
Hard not to be awed reading this. Teaser: “AGI by 2027 is strikingly plausible. GPT-2 to GPT-4 took us from ~preschooler to ~smart high-schooler abilities in 4 years. Tracing trendlines in compute (~0.5 orders of magnitude or OOMs/year), algorithmic efficiencies […]
This mirrors my (very limited) personal experience. “At least one study shows how the June version of GPT-4 is objectively worse than the version released in March on a few tasks.” ~ learn more
From the Hacker News forum: “Ask HN: Is it just me or GPT-4’s quality has significantly deteriorated lately?” My anecdotal experience aligns with this. I’ve been especially disappointed using it for code the past month or so. ~ learn more
This seems like it might be something. “Gorilla is a LLM that can provide the appropriate API calls. It is trained on three massive machine learning hub datasets: Torch Hub, TensorFlow Hub and HuggingFace. … Zero-shot Gorilla outperforms GPT-4, Chat-GPT […]
There was once a well reasoned argument for life on mars. But it was built upon faulty information, namely a visual artifact from low-resolution photos. Similarly the large language model draws reasoned conclusions as it best can, but is hampered […]
A paper published by Microsoft Research based on their experiments with an earlier version of GPT-4. If you aren’t into reading papers, that’s ok. Just scroll through and read the chat excerpts in the exhibits. You will be wowed or […]