- Product Prompts
- 😵💫 Everything AI, Everywhere, All at once
😵💫 Everything AI, Everywhere, All at once
WTH? GPT-4, Google AI, Anthropic!
Hey there 👋
In today’s Product Prompts:
GPT-4 is out, PLUS Google announces generative AI added to Google Docs, Gmail + the rest of Workspace, PLUS Anthropic opens its waitlist for Claude (as used by Notion) - BUT what does it all really mean for product people?
What, that’s not enough for one day???
I’m Martin of Product Prompts. I’ve co-founded multiple startups (fintechs) and led corporate ventures, and now offer product strategy consultancy services, including helping forward-thinking clients with the knowledge, tools and training to figure out their AI strategy for their products and internal workflows. Get in touch to learn about how I could help you.
Today has been one of those 🤯 days in AI where it’s like all the big players are showing off their latest toys to try and outshine each other. So I’m going to try and sift out the noise and highlight what it means for innovation and building products.
4️⃣ GPT-4 is here
OpenAI released its new model GPT-4 (immediately available - via phased roll-out - to ChatGPT Plus users, and a waitlist for the API). Here’s what it does:
It’s “multi-modal” - it can take an image as an input and answer your questions based on the content of that image.
💡 Example use case: Suggest some ways we could reduce friction in this user journey map.
It’s much smarter than previous GPT models.
💡Example use case: Writing API documentation
It can write code in any of the major programming languages.
💡Example use case: Companies like Cursor have already been working with GPT-4. It’s a brand new IDE built from the ground up. Beta access soon.
It has a much larger “context window”, 25,000 words of text versus the ~3,000 with GPT-3. Which means it has a much better memory so you can input/output more text.
💡Example use case: Here’s all the transcripts of our customer service calls from last month. Provide a sentiment analysis and highlight the key problems.
Announcing GPT-4, a large multimodal model, with our best-ever results on capabilities and alignment:
— OpenAI (@OpenAI)
Mar 14, 2023
🔎 Google announces its Generative AI lineup
Google launches AI across all its Workspace products: Gmail, Docs, Sheets, Slides, Images
💡Example use cases: catch me up on this 25-email long thread; rewrite this paragraph for a C-Suite audience; produce a deck including audio and video based on this one-pager.
PaLM API, an easy and safe way for developers to build on top of Google language models, and MakerSuite, a tool to jumpstart prototyping - both in private preview today. (Customers can also access these models + more via Vertex AI.)
💡Example use case: MakerSuite will let you prototype ideas, do prompt engineering, synthetic data generation and custom-model tuning. Waitlist available soon.
🤖 Anthropic opens access to its AI chatbot, 'Claude'
Two different versions are being made available:
Claude: a state-of-the-art high-performance model
Claude Instant: lighter, less expensive, faster
Early API users include Notion, Quora Poe, and DuckDuckGo.
💡Example use case: Claude seems to be more naturally conversational than CHatGPT. So this opens up another genuine competitor to OpenAI’s new ChatGPT API for businesses looking to incorporate an AI assistant.
👀 Look out for Sunday’s AI Audit.
Each week, in addition to the weekday newsletter like this one, I put myself in the shoes of a product team, ideate the generative AI opportunities and call out what I think would add the most value to the customers.
Thanks for reading, and happy prompting,
Martin, Chief Prompt Officer