Generative AI Weekly - June 20, 2024
New models, some of my own GPTs, closed tool alternatives, and more on synthetic data for market research.
This week (or two):
- I've made two GPTs I am using almost daily! Best used in voice-mode via the ChatGPT mobile app - a simple task brief writer and experiment designer.
- Anthropic releases Claude 3.5 Sonnet - beating GPT-4o benchmarks.
- "Guy Who Sucks At Being A Person Sees Huge Potential In AI"
- "Let’s start with the easiest thing I asked Claude to do: adjudicate Supreme Court cases."
- Choose a Timeless Classic [...] Gain Deeper Insight in Conversation" - Rebind (with a writeup from NYT).
- Luma Labs releases an open-for-use video generation model.
- Following Google's Illiminate from the last installment, free to use PDF to Podcast via Simon Willison.
- "Beyond Bland Output: A Few Example Prompts of How to Use ChatGPT for Brand Strategy."
- Ethan Mollick on Apple's Intelligence.
- I've mentioned synthetic data for market research a few times here - something I am extremely excited about. See Mark Ritson on the topic and a new "open-source Python package for conducting AI-powered research." Stay tuned...
- Bringing back a favorite (On Bullshit by Harry Frankfurt). "We distinguish two ways in which the models can be said to be bullshitters, and argue that they clearly meet at least one of these definitions."
- And for a paper this week - "This paper establishes a structured understanding of prompts, by assembling a taxonomy of prompting techniques and analyzing their use. We present a comprehensive vocabulary of 33 vocabulary terms, a taxonomy of 58 text-only prompting techniques, and 40 techniques for other modalities."