Rami's Readings #67 - 📢 Apple WWDC & New Model from Nvidia
The latest on AI, LLMs, Apple WWDC, Nemotron-4 340B, Financial Statement Analysis, Prompt Engineering Techniques, and more.
Welcome to Rami’s Readings #67 - a weekly digest of interesting articles, papers, videos, and X threads from my various sources across the Internet. Expect a list of reads covering AI, technology, business, culture, fashion, travel, and more. Learn about what I do at ramisayar.com/about.
Apple’s WWDC brought exciting announcements this week. In my opinion, Apple will effectively “redefine” Artificial Intelligence to mean Apple Intelligence in the eyes of most consumers - in the same way, Apple succeeded in redefining apps to mean iPhone apps. Apple will once again prove that the company that owns the distribution channel and customer interface is ultimately in control. You can see this in rumors that Apple is in talks to offer also offer Gemini and Anthropic as an alternative to ChatGPT. From Bloomberg:
Apple’s deal with OpenAI isn’t exclusive, and the iPhone maker is already discussing offering Google’s Gemini chatbot as an additional option. That agreement should be in place later this year.
Apple has also held talks with Anthropic as a potential chatbot partner, people familiar with the matter told Bloomberg earlier this year. The idea is to ultimately offer a range of AI services to users — similar to the way Apple has different search engine options in its Safari browser.
Most consumers will not venture into the Settings page where Apple will bury the option to change models. As a result, I don’t think any of the model providers (OpenAI, Anthropic, Mistral, etc) will be able to differentiate themselves and the default Apple choice will prevail with consumers. Reminiscent of search engine defaults…
Apple to ‘Pay’ OpenAI for ChatGPT Through Distribution, Not Cash
🤖 AI Reads
GPT4 Brings Moneyball Logic to Picking Stocks
Notes: A few weeks ago, I shared the paper discussed in this Bloomberg article. Remember, you heard about it here first!
NVIDIA Releases Open Synthetic Data Generation Pipeline for Training Large Language Models: Nemotron-4 340B
Notes: Huge release from Nvidia! Nemotron is “optimized” for generating synthetic datasets. More high quality data → Better models.
The Prompt Report: A Systematic Survey of Prompting Techniques
Notes: Fantastic survey of all prompt engineering techniques.
Introducing Apple’s On-Device and Server Foundation Models
Notes: Small Language Models FTW! #EdgeAI
Apple Released CoreML Optimized Image Classification and Segmentation Models for On-Device AI
Notes: #EdgeAI is happening at lightning speeds.
Steven Sinofsky’s Saturday AM Rant on LLMs on X is Worth a Read
SelfGoal: Your Language Agents Already Know How to Achieve High-level Goals
Notes: From Allen Institute for AI.
AWS Announces $230 Million Commitment for Generative AI Startups
Notes: 💰
💼 Business Reads
Twisting Apartment Towers Are Reshaping Brooklyn’s Waterfront
Notes: Grateful that not every new tower in NYC resembles a needle. Love this one in Greenpoint.
Luxury Labels Slash Prices 50% to Lure Wary Chinese Shoppers
Notes: Woah! Many of these labels are losing their status as Veblen goods.
🎨 Culture Reads
Saurabh!
Notes: Setting impossible standards for Indians kids everywhere. 😂
Yes, Everyone Really Is Sick a Lot More Often After Covid
Notes: Can confirm.
The newsletter is a bit shorter than usual this week. I officially started my sabbatical with 🎆 at a friend’s wedding in Colorado Springs. Signing off from Colorado Springs.
