GM! It’s Brett again. Every great company starts with a "what if." For Uber, it was "what if you didn't have to wait for taxis?" At Micro, our big one is: "What if your tools knew things?"

Seriously, think about it: Founders building literal rockets and robots are still manually entering data into spreadsheets and CRMs. Doesn’t make sense.

I'm talking about combining your personal context (pulled straight from your & your team’s inbox and calendar) with a global knowledge graph covering millions of companies and people - so your email client, CRM, and project management tools stop feeling dumb as hell.

I just dropped a quick product demo showcasing exactly how Micro’s global and personal knowledge graph works - powered by Perplexity and more.

Let me know in the comments there what integrations you’d want to see in Micro - I’m prioritizing waitlist spots for folks who engage.

OpenAI released GPT-4.1 family of AI models - an upgraded model with notably better coding skills, but it arrived without the usual public safety report - because GPT-4.1 isn’t a “frontier” model.

Then they also rolled out a pair of new reasoning AI models, o3 and o4-mini, designed for deeper reasoning and autonomous tool use in complex tasks. o3 is the heavy-duty version, with o4-mini a faster, smaller sibling.

These models are built to reason through multi-step tasks, use tools, and complete more complex objectives than previous releases.

But early testing raised concerns. OpenAI had previously claimed o3 could solve over 25% of problems in the tough FrontierMath benchmark, but independent evaluations from Epoch AI showed the public version scores closer to 10%. The higher number came from a more powerful internal variant with extra compute - not the model users actually get.

Hallucination rates are also up. On OpenAI’s PersonQA benchmark, o3 hallucinated 33% of the time - double that of earlier models like o1.

But that’s not all. OpenAI possibly has an internal prototype for an X-like social platform (possibly inside ChatGPT). It’s early, but if it launches, it would compete directly with X and Meta’s similar plans. The motive here appears to be data acquisition: OpenAI could get its own firehose of human interactions to train models, instead of relying on data from other platforms.

The big picture - OpenAI is moving fast, releasing new models, pushing capabilities, and expanding its footprint beyond just LLMs. But the pace comes with tradeoffs: fuzzier benchmarks, rising hallucination rates, and growing pressure on transparency. And the social platform thing is another sign the company wants tighter control over its data and feedback loops. Taken together, this week makes one thing clear - OpenAI isn’t just building models anymore, it’s shaping the ecosystem they run in.

Everyone talks about AI, but no one has the time to learn it. So, we found the easiest way to learn AI in as little time as possible: The Rundown AI.

It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.

Sign up to start learning.

Raise: Hammerspace, a data orchestration platform used by Meta and the Department of Defense, raised $100M to help AI and enterprise teams access unstructured data more efficiently.

Stat: $4.3M. That’s how much defense tech startup Theseus raised after a viral tweet about a GPS-free drone hack.

Rabbit Hole: The Inner Ring (C. S. Lewis)

The best resources we recently came across that will help you become a better founder, builder, or investor.

🚀 Plexe helps your Build ML models from a problem description.

🗣️ Aqua is an AI dictation tool that lets you talk into any text field.

⚙️ Chatbase is a platform for building & deploying AI Agents.

Simple tricks to get a ton of value with minimal effort for work and life.

Here’s a one-hour monthly ritual that pays dividends:

At the end of every month, pull up your email, DMs, or CRM. Revisit 10–15 people you’ve had some momentum with - warm leads, VCs you pitched, founders you vibed with, potential partners, ex-clients who ghosted, whatever.

Now, send out a non-salesy message. Keep it short. Share one update, ask one smart question, or send something valuable (like a deck, metric, article, or beta invite). Just signal momentum and keep the thread alive.

Optional bonus move: reach out to 3 founders who recently raised or are backed by investors on your radar. Ask for a 15-min call to swap notes - not just to ask for intros. This builds network density.

The reason you do this is because deal flow, investor interest, and serendipity all die in silence. And reviving a thread > starting cold.

Last week’s answer: If you guessed AI-generated… you were right.

🟩🟩🟩🟩⬜️⬜️ Yeah (~74%)

🟨🟨⬜️⬜️⬜️⬜️ Nope (~26%)

Here’s the original.

Micro is a CRM that doesn't feel like homework and an email client that knows what matters. Simply put, it's an all-in-one productivity tool that organizes itself so you don't have to.

I built Micro with some friends because I was tired of getting guilt tripped for not hitting inbox 0 and having to bounce around between so many tools to do my job.

That said, Homescreen subscribers can cut the line - go to micro.so/waitlist, join the waitlist, and write in “Homescreen” for Other in the “How did you hear about us” question to get early access.

If you’re interested in advertising with us, send an email over to [email protected] with the subject “Homescreen Ads”

P.S. If you're thinking about starting your own newsletter, we highly recommend beehiiv.