Gregg Luskin

Co-founder & CTO at Mothership, previously @skurt (acquired)

[Gregg Luskin @GreggLuskin]

If real world performance is anywhere near these benchmarks it’ll be a tectonic shift in AI

[Mark Zuckerberg @finkd]

Muse Spark 1.3 is rolling out today with frontier performance almost too cheap to meter. This is the biggest jump we've made so far on coding and agentic work. Try it in Muse Code and our API.

Next up 🍉 and Muse Spark open weights releases coming soon.

[Gregg Luskin @GreggLuskin]

What happens once all models hit 100% on all benchmarks.

[Gregg Luskin @GreggLuskin]

This is a great article. That said, if any of this is news or eye-opening to you, you’re already far behind the curve.

[lauren @poteto]

I'm writing a guide to pstack! Here's part one. https://t.co/4yjdWKPieE

[Gregg Luskin @GreggLuskin]

@testerlabor I double checked your math and can confirm it is 600B more

[Gregg Luskin @GreggLuskin]

Here’s a challenge for @dhh and team:

get Omarchy running on an iPad

[Gregg Luskin @GreggLuskin]

@johnhelmuth_ It’s not bizarre when you realize one was made by xAI and one was made by cursor. Grok Bot is just cursor sand rebranded.

[Gregg Luskin @GreggLuskin]

@neo_subhamoy Not really a hot take there

[Gregg Luskin @GreggLuskin]

Asked Instinct to get me a reservation (anytime in next 6 months) at a popular restaurant.

It checked, said everything was sold out, and stopped.

A real assistant would already be calling the restaurant, refreshing Resy, and watching X/IG for last-minute drops.

Still hype land, not assistant land.

[Gregg Luskin @GreggLuskin]

@solopribuilds Those two things don’t go hand in hand

[Gregg Luskin @GreggLuskin]

@gerardsans Yes it can.

[Gregg Luskin @GreggLuskin]

@jessethanley @SpaceXAI Unlikely. Very little value add to what they already have.

[Michael Truell @mntruell]

We’re sorry to see that OpenAI put out a note saying they plan to block Cursor users from accessing OpenAI models in three months.

OpenAI models serve about 5% of Cursor user traffic, and we’re speaking with the OpenAI team to resolve this.

Cursor was one of the very first users of OpenAI, we’ve worked closely with their team for years, and we’ve trusted their platform to be neutral infrastructure for our business.

[Gregg Luskin @GreggLuskin]

@theo This is not true

[Gregg Luskin @GreggLuskin]

@codeweezrd @PlanetScale It’s a hacky cursor automation right now, but it works (and saved us here!)

In a perfect world though we’d be able to set some alerting thresholds @samlambert.

Also looked at pgexporter for prometheus/grafana but we’re trying to avoid all that now that we are in PlanetScale

[Gregg Luskin @GreggLuskin]

One bad query we pushed spiked one of our @PlanetScale CPU to 100%.

No customer impact. Everything kept working.

Then our agents picked up the spike, used PlanetScale Insights MCP and Datadog MCP to trace the logic across services, and fixed it.

What a world we live in.

[Gregg Luskin @GreggLuskin]

TPS is the understated bottleneck to flow state.

Running 10+ agents in parallel is normal now. That's not the problem. The problem is working through something deep and complex then sitting there for 20 minutes while it thinks.

This will be a solved problem in a year or two. Until then it still kills that small percentage of critical work.

[Gregg Luskin @GreggLuskin]

@solopribuilds That’s not a hot take that’s called statistics.