This month, instead of sharing a think piece of my own, I am sharing with you think pieces / conversations from some of the best in the business. Check them out below in this late-month edition of Future Telescope.
1. Ben Evans on the utility of generative AI
Benedict Evans, a long standing authority on what is up in the world of technology, has shared this deep think piece on the actual real world use cases of generative AI as it stands today. He gives an excellent analogy at the outset - just as PCs were not of much use to the public at large till tools like VisiCalc, the predecessor to Excel came about, and how tabulation software was of more use to accountants than it was to lawyers who needed a word processor or graphic designers who needed Photoshop. We are yet to see how LLMs like ChatGPT can help larger parts of the economy other than just students writing their papers or developers writing code. There will need to be a better use case that makes these products stick, and Ben has some compelling thoughts here. Check it out!
2. Ezra Klein interviews Dario Amodei, the CEO of Anthropic, the company behind “Claude”
Anthropic, the company behind the recently much acclaimed “Claude 3” chatbot, was founded by OpenAI alumni. One of them, the CEO Amodei, sat for a very well thought out discussion. He speaks in this conversation about the “scaling laws”, where technological development in the AI space scales exponentially with time and investment. This growth of technological capability he says is on a smooth curve, but the realization of such technology in the real world is “spiky”. In his words:
And then there’s a spilling over of that process into the public sphere. And the spilling over looks very spiky. It looks like it’s happening all of a sudden. It looks like it comes out of nowhere.
This is actually a very important thing to keep in mind as you dive into the world of AI, dear reader. I remember seeing the development of GPT 1 through 3 from 2018 to 2020, and didn’t feel as surprised when ChatGPT came out, as I saw it as a new wrapper around an existing technology. Now you are here in this world with me, learning about the developments in this space, and thus when the next bomb drops in the AI world, you may end up being way less surprised by it than you were by ChatGPT!
3. Dwarkesh Patel interviews Mark Zuckerberg
Many of us may remember the worst quarter for Meta that transpired in Feb 2022 due to its metaverse investments. Shares of the company slid 22% in just that month. But few of us may know that most of the company’s costs in that quarter were actually attributable to beefy Nvidia GPUs that it purchased to power its AI efforts like Llama and the Rayban Meta glasses. This power move by Zuckerberg also bled into Meta’s adtech tools called “Meta Advantage”, offering better ad targeting in spite of new restrictions in customer profiling across iOS and Android.
Now, Zuck is back to talk more about AI from the recently released Llama 3 to his thoughts on AGI and how he thinks we may be headed for another AI winter, because of the stagnation in innovation in the generative AI space.
Good chat, check it out!
4. Ethan Mollick on using LLMs to innovate
While an LLM is best at guessing the probability of one word to follow another, it is also fertile ground to innovate and come up with ideas from scratch.
, a Wharton professor, is great at finding such opportunities of innovation through AI tools, and his substack is a great place to see his experiments with various AI tools. This is a great thinkpiece written as a companion to one of his upcoming research papers, and I recommend you to check it out!5. Aspiring for Intelligence analyzes the AI stack of the Big 5
and from talk about the big 5 cloud providers Microsoft, Amazon, Google, Snowflake, and Databricks, and the AI stacks they have built across model, infrastructure, and application layers. I got super excited about Amazon’s upcoming model, and to see what specific problems they try to solve apart from being just another chatbot model. I think given the enterprise focus of a potential Amazon model, low cost per token, action oriented agents, and large context windows similar to Claude 3 and Gemini 1.5 will be table stakes. What gets added on top of these table stakes will be exciting to see.6. MKT1 Newsletter on AI usecases for marketers beyond ChatGPT
While the think piece by Ben Evans stresses the use-case issues with generative AI, this post by
dives deep into the tools available on the market today for a specific audience - marketers, and specifically marketers who work in the tech space. There is an automation agent similar to Devin the coding agent, a presentation maker, and a video asset creator in the mix. All in all a good look at the current state of tools and a good place to get some ideas on how you can improve your workflows if you are a marketer.There is a lot happening in this world of AI, and I hope Future Telescope is a place where you get to learn about this beautiful world in depth.
That’s it for this edition, see you next month!