ZD 25.12: Learn AI Like a Pro II
How to keep up with AI.
In the Issue: Guidance on how to keep up with fast-moving AI through a curated selection of newsletters, video channels and communities.
The Distilled Spirit
Tech Change, By AI
🚫 No Prompt Prompting FTW ()
Newer, faster AI engines are changing prompting techniques. One thing that works very well, especially for the effort, is just pasting in the content you want AI to work with sans prompt. This often triggers a largely correct response from the LLM, greatly reducing barriers to entry to AI usage. Don’t overthink it.
🍎 Apple Suffers Post-Traumatic Growth ()
Apple Intelligence just had a major failure to launch. The AI revolution is giving Apple fits. Liberty argues that Apple should become a neutral platform where AI providers could compete. To become the Switzerland of AI so to speak.
🎧 Joanna Stern Vibe Codes (Wall Street Journal)
In case you missed it, Vibe Coding is hot these days. So hot that Joanna Stern checks cursor and reports on her experience. We can all build bedtime games in 2025.
😁 How Technology is Making us More Human ()
AI is making the world better in a lot of ways. AI powers medical breakthroughs; it empowering employees to be owners and it opens access to unlimited knowledge and entertainment. The future is bright if you know how to grab it.
Cultural Shifts
💬 Conversation is the new Unit of Culture (
)Conversations are eating the world, one TikTok and NotebookLM generated Podcast at a time. Kyle Chayka interviews writer Tope Forlarin to explore and flesh out how the activity and authenticity lends authority and weight to the medium.
🍀 Putting the Irish in Pubs Everywhere (Smithsonian Magazine)
Irish pubs are everywhere, in no small part due to the Irish Pub Company. This Dublin-based outfit has designed upwards of two thousand pubs in one hundred counties around the world. They are not pubs in a box — each is custom designed. The company sells Irish pubs as a vibe.
Other Interests
⛵ The Rise and Fall of the Hanse ()
The Hanseatic League was an ambiguous beast. It was neither state nor company entirely. Not a kingdom, no central decision making body. It might not have even been a formal League. It did make it’s mark on the economy and laws of much of northern Europe in ways that are felt today.
📚 Will Storr says You Are a Story ()
Bestselling author Will Storr has launched a substack focusing on storytelling. Stay tuned.
🚗 50 Years of BMW Car Art (BMW via Wallpaper* Magazine)
I love BMWS. I love art. The two together are beautiful and you should enjoy them.
The sky hook was an artful, un-blockable shot invented by Lew Alcindor — later known as Kareem Abdul-Jabbar. But when did he invent it?
👠 The Cinderalla we almost Got ()
McNeese State faltered against a much better Purdue team so we will no longer get to enjoy their walkout and the rise of Amir Khan.
AI Generated Musical Coda
Please enjoy this week’s AI generated musical coda featuring an east coast underground hip hop theme.
Learn AI Like a Pro in 2025
A year ago, when I wrote the last version of ZD 24.04: Learn AI Like a Pro, it was a much different landscape. ChatGPT 4 was still pretty fresh and premium-only. Context windows were comparatively tiny; you were limited on how much content your AI could use. Multimodal capabilities — the enhanced versions we have today that can see and hear — were months off. Function calling and tool use were lab features at best. Fast, cheap flash and mini versions of models were yet to come into their own. We were still writing our own Chain of Thought prompts; models did not yet think it out for us. Web search was Perplexity or bust. There was no such thing as Deep Research mode.
AI has moved fast. We have GPT-4o-level models on the free tier from most providers. Context windows are giant. Function calling, flash and mini models, and vibe coding are creating a wave of AI-powered apps. Multimodal tools let those apps see, hear, and speak. Your models think much more effectively — and even explain their reasoning. For example, they’ll often walk you through the steps they took to reach a conclusion or suggest how to improve your prompt for better results. You can ask for Deep Research from a few providers, and almost everyone has web search.
To keep up with the field, you have to pay attention to a lot of signals. Here is a heavily curated list of creators and channels to follow to keep up with the breakneck pace of AI today.
🕵️♀️Tracking AI News
Keeping up with AI can feel overwhelming — but the right sources make it manageable. Here are a few standout newsletters that offer reliable, well-curated insights at just the right pace.
I need a fast summary of the AI news of the day
TLDR ingests the technology firehose and pares it down to a list of stories you can digest between metro stops. All of their publications are worth the free price of entry if you have the interest. For the purposes of this publication, check out TLDR AI for the best daily AI news summary out there.
I need a daily that captures the big stuff but does good on AI software and tools too
Noah and Grant write The Neuron, a daily publication that does a good job curating stories and tools. They list and describe a dozen or more different apps that are often worth investigating, as well as track LLM updates — making for a good daily 2-page update in your inbox every weekday.
I need a monthly AI roundup, the day-to-day is too much!
Nathan Beniach, founder of Air Street Capital, spends a lot of time investing in AI and life science technologies. He writes a monthly
that ties up all of the important stories in the fast-moving field. Great for keeping up without getting sucked too far in.China is where it’s at — how can I keep up in English?
China is a very important part of the AI story. Jordan Schneider and his team at
are doing an enlightened job of staying on top of developments there and making it accessible to English-language readers. They provide a unique perspective worth tracking in the worldwide AI race.What if a rationalist read all of AI thought on X and curated it down to a lengthy yet amazing summary that showed up in your inbox every week?
Zvi Moshowitz is a rationalist writer who does an amazing job curating all things AI at
. He writes a weekly AI roundup that is long but well-structured and easy to scan. He also publishes timely briefings on hot topics in AI. The writing is sharp and the wit quick.⚒ Learning How to Apply AI
Are you looking to apply AI to your work or life? Look at these newsletters for a good idea about how to get there.
I want to understand how to work with AI in a grand sense.
Professor and author Ethan Mollick is a leading thinker on working with AI. Co-Intelligence is a must-read for workplace readers and strongly recommended for anyone curious about the best ways to work with AI. The book is an extension of his Substack,
, which is the best way to keep up on how to work with AI at a high level in 2025. His tentpole articles — like the very recent thesis about cybernetic teammates — are worthy of study.I need tips on tools to use and tricks about how to use them.
Jeremy Caplan is an educator and journalist who loves creative tools that make life better. He writes a Substack called
that drops a weekly guide on tools you can use immediately. If you’re not subscribed, check out pieces like his AI productivity toolkit to see what you’re missing.I want to know how I can use AI for my knowledge work and productivity.
Every is tough to describe — it covers most facets of AI. Writers like Dan Shipper and Nathan Baschez explore everything from prompt engineering to AI-enabled productivity. A way to think of it is as a more practical version of Ethan Mollick’s work — a strong subscription if you’re looking to lead workplace AI in a hands-on way.
🔥 Living on the Bleeding Edge
Are you trying to keep up with the bleeding edge? These are the two resources I use to stay on top of developments without getting overwhelmed.
I want to understand the bleeding edge of practical client and LLM applications.
Simon Willison is a notable software developer who co-created Django and helped build Eventbrite. He’s now focused on open-source data journalism and AI. To understand how cutting-edge LLM developments become usable, follow his frequent posts and weekly roundup at
.I want to read all of the AI news on X, Reddit, and specific AI Discords — without having to subscribe to all those things.
The Smol AI company publishes AI News, an AI-generated summary that covers an amazing breadth and depth of AI channels. It includes a categorized recap of AI on X, a curated selection of Reddit and Discord highlights, and much more. It’s the best way to feel the pulse of the field without personally diving in.
Worthwhile Video Channels
Do you prefer to learn via video? Check out these channels for some good ways to watch and learn about AI.
I want to understand the big industry and product picture and look into how this will change business. I want to get AI fluency.
Nate B. Jones publishes across multiple platforms. He started on TikTok, has a solid Substack, but is strongest on YouTube. He shares insightful takes on AI developments and their impact on products, companies, and the future of work on a daily basis. It is a pretty functional five minute synopsis about the biggest thing moving AI that day in most cases.
I am a product manager looking to leverage AI.
Peter Yang has become a standout voice in the space lately. His handy guide on which AI to use is a great reference. His YouTube channel, originally focused on product management, has recently shifted into strong AI coverage. The recent Cursor tutorial is a great example. His Substack is also worth a look.
I want to learn how to use AI for things I actually use and do.
Jeff Su shares practical tips on cloud apps and productivity tools on YouTube, with a clear focus on how AI enhances day-to-day workflows. His content is approachable and actionable — perfect for anyone looking to make smarter use of AI in their daily routine. Check out his videos on AI productivity hacks and top AI-powered extensions for Google Workspace as great starting points.
I am a developer looking to learn how to create AI-powered applications.
DeepLearning.ai offers short video courses that are free and great for getting started. These are excellent for learning how to build AI apps or understand what’s possible.
🗯Communities and Chats
Looking for a supportive AI learning community? These are two clubs I have found to be worth joining.
I am a non-technical person looking to learn how to build in AI.
Ben Tossell started Ben’s Bites as a newsletter and grew it into a learning platform. A membership gets you access to tutorials, workshops, and a supportive community.
I am a leader charged with implementing AI and need to talk to other leaders.
Daan van Rossum’s Lead with AI PRO community is built for executives and team leaders. It’s a smart way to network, keep up with developments, and find practical support.
What’s the best way to keep up with my important AI vendors? How can I get help with the API?
Discord is the tool many AI vendors use for support and announcements. It’s a direct line to the people building your tools. Join the OpenAI, Anthropic, and Perplexity Discords to stay in the loop.
❓ Bonus: Platform FAQ, Spring 2025 Edition
Should I pay for a chatbot? Which chatbot should I pay for?
Free versions of leading chatbots are getting better. But they’re still limited — especially on volume and integration. If you can spare $20/month, it’s a smart investment. Paid versions unlock multimodal capabilities (text, image, voice), and that alone is game-changing.
ChatGPT is still the leading choice — strong engine, strong product. Google is a close second, especially for Android users. For more, see Peter Yang’s comparison of the major options.
DALL·E 3 isn’t cutting it — what are better AI art tools?
DALL·E 3 (included with ChatGPT) is a good starting point. But if you want more control or features, try Leonardo.ai or the new Midjourney web app. ZD 25.06: Presenting with AI also covers graphical and presentation development AI options.
I have a bunch of info I need to synthesize — is there something quick, easy, and free?
NotebookLM is a wonderful tool. Google gives you lots of functionality for free — you can upload up to 50 documents per repository. Then it helps you draw insights and even generate a podcast explaining the key ideas.
🎁 Conclusion: Learning AI in 2025
The AI world is wide, wild, and moving faster than ever. Staying current doesn’t mean reading every paper or testing every model — it means knowing where to look, who to follow, and what tools make sense for your goals. Hopefully this guide gives you some good options to explore or apply in your work or personal projects. If you have some favorite publications I don't mention here, please share them. I am always looking for new resources and this community would undoubtedly be interested as well.
The Look
AI generated reddit profile summaries are hot.
Did you enjoy reading this post? Hit the ♥ button above or below because it helps more people discover great Substacks like this one and it helps train your algorithm to get you more posts you like. Please share here or in your networks to help us grow!
Hey, thanks for the link, Wyatt! 💚 🥃