Welcome to your weekly Brew & AI
Each week, I’ll share how to make sense of AI - no jargon, no hype, just simple insights you can actually use.
P.S. - To ensure you receive this newsletter, please move it to the primary tab and set it for all future communications. Or just add this emailid as a known contact. I wish AI could solve this for me.
Grab your coffee - let’s dive in. 👇
☕ AI in the news
Claude quietly goes ultra‑long: 1M‑token context is now GA
Anthropic just made a 1 million token context window generally available for its latest Claude Opus and Sonnet models, so you can stuff entire codebases, wikis, and multi‑year email archives into a single session. Requests over 200k tokens now work without any beta headers for these models, making “long‑context” feel like a default capability rather than a special research feature.
It’s a strong signal that the next competitive frontier is not just model IQ, but how much of your world an AI can see at once.
To put this into perspective, 1M tokens would get you roughly the entire collected works of Shakespeare (900,000 words) with room to spare, 10,000 emails (at ~100 words each), or 30,000–50,000 lines of code.
Chip wars heat up: AI is coming to everything
Nvidia, AMD, and Samsung are all racing to put faster AI chips into the devices we already use every day.
Nvidia is building new hardware for giant AI models in data centers, while AMD is pushing laptop chips that make AI tools feel instant without always needing the cloud. Samsung, meanwhile, wants hundreds of millions of its phones to run Gemini on‑device, which basically means your “AI assistant” will just live inside your phone, not a flashy new gadget.
Yann LeCun’s new startup raises $1B to chase “world models,” not just chatbots
Yann LeCun’s (former chief AI scientist at Meta) new company, AMI Labs, quietly raised about $1.03B in seed funding - backed by Nvidia and Bezos’s family office - to build so‑called “world models.” Instead of another giant LLM, they’re focused on AI that learns the physics and structure of the real world, targeting robotics and industrial automation where current chatbots often fall apart.
The signal: the next AI wave may be less about writing emails and more about machines that can actually understand and act in messy, physical environments.
☕ Trending on social
In What’s Trending on Social, I’ll surface the most interesting AI takes, memes, debates, and hot posts bubbling up across X, LinkedIn, and beyond - so you don’t have to doomscroll to stay in the loop.
🧑💻 Elon Musk’s xAI just poached two of Cursor’s top leaders, Andrew Milich and Jason Ginsberg, to rebuild its struggling AI coding tools. Both were central to Cursor’s rise as one of the hottest AI coding assistants, and they’ll now report directly to Musk, highlighting how intense the talent war has become in AI developer tools.
🗺 Google is rolling out a new 3D “Immersive Navigation” mode in Maps that makes driving directions look much closer to real life. When you navigate, the map now shows 3D buildings, overpasses, and terrain, highlights lanes, crosswalks, traffic lights, and stop signs, and uses smart zooms plus updated voice guidance so turns and tricky junctions are easier to understand at a glance.
📊 Andrew Curran joked that we’re “cooked” as Anthropic and OpenAI keep shipping faster than anyone can keep up, even if you spend all day in this stuff. It’s a very relatable “I follow AI for a living and still feel behind” vibe that pretty much sums up how the whole space feels right now.
☕ AI workflow of the week
Each week, I’ll spotlight one simple, battle‑tested way to plug AI into your actual work - not just a prompt, but the mini‑workflow around it. Think of it as a copy‑paste system you can run in 10–15 minutes to save yourself an hour.
Do a complete competitor teardown with an LLM
Who this is for: marketers, PMs, founders, strategy folks.
Problem: You want to understand a competitor deeply (positioning, features, pricing, ICP, go‑to‑market), but doing it properly means trawling through their site, docs, blog, X, and review sites. It’s a few hours of work, so you rarely do it properly.
How to run it (30–45 minutes):
1. Feed the model raw evidence, not vibes. Paste in chunks: website copy, pricing table, feature pages, and a few real user reviews. For very long input, drop everything into a doc and use a long‑context model in one shot.
2. Ask for a structured teardown, not a summary. Prompt it to produce a 1–2 page report with sections:
“Product & feature overview” (what they actually do)
“Target customer & use cases”
“Positioning & messaging” (how they talk about themselves vs. category)
“Pricing & packaging” (tiers, key trade‑offs)
“Strengths, weaknesses, opportunities, threats” (SWOT)
3. Generate role‑specific views. Run follow‑up prompts like:
“From a marketer’s POV, what are 5 angles they’re leaning on that we’re not?”
“From a PM’s POV, what features or UX decisions are real differentiators vs. table stakes?”
“From a founder’s POV, where are they most vulnerable in the next 12 months?”
4. Turn it into action. Finally, ask: “Based on this teardown, give me a prioritized list of 5 experiments we should run (messaging tests, feature ideas, content pieces), with 1–2 lines of rationale each.”☕ Try this out - prompts
Here, I’ll share some interesting prompts that I’ve either used myself or come across on social media.
This week - Turn an AI model into your slightly unhinged “make my job cooler” coach.
You are my chaotic-good work coach.
Ask me 5 quick questions about my job, then propose 3 specific, slightly ambitious ways I could use AI in the next 30 days to look obviously more effective and interesting at work.
For each idea, give:
- a fun title,
- what I actually do,
- the tools I should try,
- and the “brag line” I can use when I talk about it.☕ Weekly Reads
Each week in Weekly Reads, I’ll share a short list of the most insightful, no‑nonsense AI articles I’ve found across the internet - pieces that explain what’s happening, why it matters, and how you can actually use it.
☕ Resources
If you missed my writing, here’s everything I’ve done - bite‑sized, practical, and very AI‑obsessed. Think of this as the “in case you blinked” shelf for my recent posts, threads, and experiments.
✍ Blogs - In-depth articles and insights to expand your AI knowledge
🛠 Tools - Discover powerful AI tools to enhance your workflow
💭 Tips - Practical tips and tricks to make your AI journey smoother
💛 P.S.
That’s it for this week’s brew.
I’d love to hear what you think - what you liked, what could be better, or what you’d love to see next.
Just hit reply - I read every message over my morning coffee ☕.
Brew & AI
Making AI simple, one sip at a time