FINALLY! The new open source LLM released by @databricks, Dolly 2.0, has a *clear* answer to my question on how to get rich quickly.

It succeeded where ALL the others failed (see the other answers in the 🧵).

Thanks @RamAnanth29 for putting together a HF space to find the truth https://t.co/zspMhd2XdQ

@GaryMarcus All of this is very euphoric and very, very premature. Imagine what happens after next week when OpenAI shows GPT-4 plugins at TED.

This week, as usual, I have a few people to thank for their *involuntary* contribution to the Free Edition of Synthetic Work: @remixeverything, @emostaque, @Ninetail_foxQ, @paulg, @asanwal, @emollick, @vijaybolina, @openai, @maurice_jks (& team), and @Kyle_L_Wiggers.

1/ This week I’m introducing a new section in the Splendid Edition of Synthetic Work. It’s called “Prompting” and it’s all about getting the most out of your interaction with GPT-4 (and, in future, other generative AIs) thanks to a few advanced techniques to write your prompt.

2/ Here’s what’s in it.

To understand how prompting techniques influence the outcome of our requests, I showed the comparison between a standard prompt and a prompt modified with these techniques. https://t.co/thwbI7D3re

The new issue of Synthetic Work is out, people.

This week’s Splendid Edition is titled Issue #8 – The Harbinger of Change

Read it here: https://t.co/n3NZww7GFi https://t.co/SkjvUL6AMj

The new issue of Synthetic Work is out, people.

This week’s Free Edition is titled Issue #8 – The Best Funeral of the Year

Read it here: https://t.co/VOBds9ekEe https://t.co/qGSB6MCnuc

@_ScottCondron @weights_biases @StabilityAI Scott, LangChain’s first release was 6 months ago…
When you say “5y of experience” what kind of years are you talking about? What planet should we use for the reference measurement?

See you next week.

Go @ispace_inc

@mpshanahan @DrRyanBurnell The people that need to read this don’t have a subscription to Science.