🏠 Claude Opus 4.8 just beat GPT-5.5... mostly

GA! It's Brett.

Happy first day of June!

We’re throwing a big rooftop party for NY Tech Week on Wednesday! You should come out and say hi!

RSVP to the Startup Soiree here

Alsoooo on the Micro front, we just partnered with Slash to give Slash customers a masssssive discount. Sign into Slash or go here to redeem.

And most importantly we have a HUGEEE LAUNCH on Weds.

You can get a sneak peak of what we’re launching here or by clicking the button below - you’ll also get 1 month free + 10K AI credits.

No fancy launch video compares to a raw demo of your product doing amazing things.

Farza dropped one this weekend — controlling his entire computer with just his voice. No hands. One take. Showing his screen live while telling the AI to make him money. 9K views and climbing.

The genius of it breaks down like this: everyone wants to see how much money people make, so showing a live Stripe dashboard immediately creates stakes. Telling AI to make you more money is something everyone dreams about. And the low-effort nature — no edits, no animation, just a real product doing something real — multiplies the credibility.

Founders are tempted to do fancy animations on launch videos. The truth is if your product is really magical, you should show the actual product. It's more trustworthy.

The easiest way to make a legendary demo: jump on a powerful new AI capability the moment it drops and show what your product can do with it. Farza did exactly that. The result is the kind of video that makes people stop scrolling.

We're doing a Micro launch video this week. Watch this space.

What’s changing in how companies expand globally?

Hiring internationally used to mean opening entities, navigating months of legal setup, and building local infrastructure before making a single hire.

That model is starting to shift.

More companies are using EOR not just as a temporary solution, but as a strategic way to access talent faster, test new markets with less risk, and scale globally without adding operational complexity too early.

But the biggest change may not be the hiring model itself. It’s how companies think about expansion.

Instead of building infrastructure first and hiring second, many teams are now hiring where the best talent already exists — and building strategy around that reality.

Oyster’s Strategic EOR Whitepaper explores how modern companies are using EOR to scale internationally, where the model works best, and why the global expansion playbook is evolving faster than most leaders realize.

Download the report

The pace is the story. Anthropic is shipping frontier model updates faster than most companies ship product releases. Opus 4.8 dropped last week — six weeks after 4.7, which itself came six weeks after 4.6.

The headline feature: Dynamic Workflows. Instead of one agent working through a task sequentially, Opus 4.8 can spin up coordinated swarms of subagents running in parallel — one researching, one writing, one checking, one executing — all orchestrated by a single prompt. The result on SWE-Bench Pro, the most rigorous software engineering benchmark: Opus 4.8 beats GPT-5.5 by more than 10 points.

The underreported feature: uncertainty calibration. Opus 4.8 is four times more honest about what it doesn't know compared to 4.7. That sounds like a footnote. For enterprise use cases — legal, finance, medical, anything where a confident wrong answer is worse than no answer — it's the most important improvement in the release. A model that knows when to say "I'm not sure" is worth more in production than a model that's always confident.

The practical implication: if you're using Claude for anything that involves multi-step research, complex code, or agentic workflows and you haven't upgraded to Opus 4.8, you're leaving real capability on the table. The Dynamic Workflows feature alone changes what's possible in a single prompt.

The bigger picture: KPMG just deployed Claude across 276,000 employees. OpenAI Codex is autonomously controlling Windows PCs for testing and bug hunting. The gap between what the models can do and what most operators are actually using them for has never been wider. Opus 4.8 widens it further.

Read the full release notes.

4 things on my mind this week:

@Vivek4real_ on Larry Ellison's take on the only moat left in AI

Black cat. White fence. Looks like someone's phone shot from a moving car.

Last week's poll: Real. 66.7% of you got it right — voted "Nope" correctly. The 33.3% who said AI were fooled by the dramatic contrast and skin texture processing, which looks exactly like what AI generates when you say "hyperrealistic portrait." The eyes especially. Real life keeps winning this contest.

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.

LHV is a community-driven, early-stage venture fund backing ambitious founders at the earliest stages.

If you’re building tools that help founders, investors and other professionals who want to be 100x more productive with 100x less work, and you’re interested in advertising with us, send an email over to [email protected] with the subject “Homescreen Ads”