Sean Gartland

Product @offerlabHQ | Founder, Pocketsomm | Wine enthusiast (WSET 3) | Ex-ESPN, Rotten Tomatoes

[Sean Gartland @SeanKaveh]

Most AI tools now split into modes. Chat over here, code over there. There's also a big push toward managed agents right now, with Claude and nearly every other platform racing to ship orchestration frameworks and fleets of agents.

I wanted to see how far I could get staying lightweight instead. So I run one persistent agent I call Free agent. It's a quarterback that holds deep context on our whole app and delegates from there, auditing our Linear cycles, sweeping QA feedback for small fixes, cleaning up interaction details, and spinning out other agents when a job needs its own focus. It even drafted a pitch deck one-pager last week, then handed off to a second agent that built the full deck in @Replit.

Not claiming it's the right answer, just what's working for me so far. Curious where others land: heavier managed-agent setups, or lightweight single orchestrator like this?

[Sean Gartland @SeanKaveh]

The unofficial rule for the last two years was simple: use as much AI as you possibly can. Burn tokens, poke at everything, find the edges.

That was the right call.

@RampLabs Router is the signal we're turning the corner, and what I find sharp about it is the bet underneath. The router isn't really the product. It's the foundation everything else gets built on, the layer that quietly decides which model handles which request. Own that decision and you own a lot.

Here's the part most takes are missing: this only looks like a cost play on the surface. What Ramp is really betting on is the curve. Lower cost, more efficient models are improving fast, and the gap between the cheap option and the frontier keeps closing. So routing every request to only the model it actually needs isn't just cheaper today, it gets more viable every single month.

Yes, there's some short-term fidelity risk in sending the wrong request to the wrong model. But that risk is shrinking quickly, and building the routing muscle now means you're positioned to ride the improvement curve instead of chasing it later.

That's the right bet to make. The interesting question is which teams build the judgment to route well while the models catch up. Smart move by the @tryramp team.

[Ramp Labs @RampLabs]

We’re launching Ramp Router, our LLM gateway.

It processes trillions of tokens a day across Ramp’s external products and internal AI workflows, giving us one place to manage reliability, latency, and cost.

Reserve access today.

[Sean Gartland @SeanKaveh]

@DougDribbl88 @JJWatt @JJWatt is an absolute legend. What a class act!

[Sean Gartland @SeanKaveh]

These are clearly designed for single floor living, is that a broad enough market? In California almost everyone is on 2-3 floors, so I can imagine this pricey Isaac crashing down when it attempts to empty the dryer and bring the clothes downstairs 😭.

[Weave Robotics @weaverobotics]

Today, we’re launching our home robot Isaac 1.

Isaac 1 deliveries will begin this fall.

Order yours below.

[Sean Gartland @SeanKaveh]

I see this at least 3 times a day. 😂

[Peter Yang @petergyang]

what does this even mean

[Sean Gartland @SeanKaveh]

@claudeai: You're out of usage credits.
Me: I literally have $90ish dollars worth left, what are you talking about? 🤦‍♂️

Keeps forcing me to reload with more credits even though I have a ton left, what gives?

[Sean Gartland @SeanKaveh]

@petergyang Yeah I think that’s code for: Disable it for everyone. A ridiculous requirement the government clearly knew couldn’t be precisely enforced.

[Sean Gartland @SeanKaveh]

This tells you everything you need to know.

[Ryan Carson @ryancarson]

If you have ANY private repos with plain text secrets or sensitive documents/architectures, immediately rotate your secrets

[Sean Gartland @SeanKaveh]

@aakashgupta Better solution (something Apple is also very bad at): Wireless charging pad, most use mouse pads anyways, make a beautiful one that charges the mouse while it’s on it and make it completely port less.

[claire vo 🖤 @clairevo]

Tough day for folks (DM me if I can help!) but if I found myself suddenly laid off from a company that cited AI as a cause, this is what I’d do:

- download codex and Claude/code
- say: “this was my job and how I spent my day, how can you help me automate it w skills”
- push a dozen of those skills to GitHub
- open up Claude design and make a portfolio site, with an “agent” per skill explaining how you built it, what tools it interacts with, etc.” give to lovable or v0 or whatever to publish
- post that site and link to GitHub on LinkedIn
- search “ai for <job>” and try all the new startups, form an opinion, message their founders
- try something scary like openclaw, form an opinion
- take a course in tactical AI in your field
- build, share, build

The gap in AI adoption is getting bigger. Start reskilling now while it’s early.

The time is now.

[Sean Gartland @SeanKaveh]

@mattpocockuk 🤣🤣

[Sean Gartland @SeanKaveh]

Since we’re all using AI here to write for us:

Some of the chemistry is directionally true but flattened. Allicin does degrade, lecithin does emulsify, glutamates do stack — but the framing is too clean. “Caesar activates umami receptors closer to a steak than a plate of greens” is the kind of sentence that sounds like a fact and isn’t really one. The Cardini origin story is real but the “he assembled it tableside because the dressing degrades” causal claim is almost certainly invented post-hoc — the actual story is he was running low on ingredients and improvised, not that he was solving a chemistry problem.

I’m only giving you a hard time, it’s all love @aakashgupta 😜.

[Sean Gartland @SeanKaveh]

In 2026, it should be illegal for any processor to require filling out your credit card info on a physical paper and mailing it back as the only form of payment.

Not having secure online billing is unacceptable IMO. Completely ridiculous!

That includes medical billing, namely.

[Sean Gartland @SeanKaveh]

Very cool, definitely feels like something we should expect in the future.

[Zain Shah @zan2434]

Imagine every pixel on your screen, streamed live directly from a model. No HTML, no layout engine, no code. Just exactly what you want to see.

@eddiejiao_obj, @drewocarr and I built a prototype to see how this could actually work, and set out to make it real. We're calling it Flipbook. (1/5)

[Sean Gartland @SeanKaveh]

Claude just added connectors for Instacart, Resy, Booking, Spotify and more.

Here's what most people are missing: apps are becoming services that AI orchestrates, not destinations users visit.

In e-commerce, this changes everything. The next generation of brand partnerships won't be "collab drops" on Instagram. They'll be two brands showing up together inside an AI recommendation because their products actually complement each other.

That's the future we're building at @OfferLabHQ .

[Claude @claudeai]

Claude can now connect to more of the apps you use outside of work, including @Tripadvisor, @bookingcom, @resy, @Instacart, @Spotify, @audible_com, @AllTrails, @thumbtack, Intuit @turbotax, and more.

[Sean Gartland @SeanKaveh]

@KingJames @odysseymovie 🔥🔥🔥🔥🔥

[Akshay Kothari @akothari]

You can use AI to cut costs. You can also use AI to raise your ambition.

Too many people focus on the first. More people should be thinking about the second.

[Felix Lee @felixleezd]

Claude Code is absolutely incredible but have you tried going outside?

[Sean Gartland @SeanKaveh]

“Just ship an AI agent” they said.
*builds sandboxed execution environment for 6 weeks*

Claude Managed Agents abstracts away the months of infrastructure work that has kept most enterprises from shipping agents in prod.

The real unlock here isn’t speed to prototype. It’s speed to production.

[Claude @claudeai]

Build and deploy your agents through the Claude Console, Claude Code, or our new CLI: https://t.co/E9xQ7xd4rG

Read more on the blog: https://t.co/omWjJ4fK88