←
Chris Halaska10,753 posts
Chris Halaska avatar

Chris Halaska

@chalaska

Design engineering + agentic systems 20+ years in products, ex. @Google

◷ Joined April 2009▧ 941 Photos and videos
Chris Halaska avatar
Chris Halaska@chalaska·2h

I run the studio's operations on Grok Bot. It's set up as a small team of AI agents, each with its own role and memory: Chief of staff: runs my day, triages inbox and Slack, keeps the pipeline current. Sales: briefs me before discovery calls and debriefs after. Design: tracks progress and feedback across live client projects. Growth: manages my posting schedule and flags followers who look like leads. Accounts: client onboarding and invoicing. Analytics: watches site performance and search. At 8am I get a brief. What changed overnight, what needs me, and a plan for the day. Nothing goes out without me. Every email is a draft I approve and send myself. It's freed me up to spend far more time on the work and the rest of running the studio. A lot of this used to swallow me for hours at a time.

Chris Halaska avatar
Chris Halaska@chalaska·12h

In my second month running the studio we had 13 clients at once. Great for revenue. Terrible for going deep on any of them. Now we keep it to a handful at a time, so each one gets real focus.

Chris Halaska avatar
Chris Halaska@chalaska·15h

The cheapest fix for a product that looks AI-made is usually the type. Before colour, before layout, before a redesign. A venture studio founder showed me one of their AI-built sites last month: "These are your typical AI fonts that are being generated. Even something as stupid as just updating the font makes a big difference." Defaults give it away first. It's a shame too. AI has killed a lot of good serif fonts.

Chris Halaska avatar
Chris Halaska@chalaska·Nov 17

Currently designing an interface that’s designed not to be used. MCP is becoming the primary use case.

Chris Halaska avatar
Chris Halaska@chalaska·Nov 17

I think the best early product team right now is two people. A founder who knows the market. And one person who can take an idea from brief to shipped screen without a handover. That second seat used to be three hires.

Chris Halaska avatar
Chris Halaska@chalaska·Nov 16

A founder asked me last month whether to buy design as sprints, weekly or on a retainer. He'd had proposals for all three. I said retainer. This was months of work to get right, not a couple of sprints. And his product needed cutting back, which sprints work against. More features means more sprints, so nobody on either side is in a hurry to remove anything. A big part of my job is saying no and taking things out. Easier when next month isn't riding on the feature list. If the work has a clear end, a fixed project makes more sense. This one didn't.

Chris Halaska avatar
Chris Halaska@chalaska·Nov 16

Discovery, definition and design used to run one after the other on my projects. Now they run together. I can prototype an idea in code while we're still working out the direction, so a founder is reacting to something real in the first week. Discover and define. Design. Deliver. Same steps. A lot less waiting between them.

Chris Halaska avatar
Chris Halaska@chalaska·Nov 16

Adrian’s right on this. My part of the work on a live product is shorter than it’s ever been. A round of direction, review what comes back, another round. But that direction took years to build. Knowing what to ask for, and what to throw out. The price of great work hasn’t changed. What you’re paying for has.

Last year I said designers would either scale their output or drop their prices. That's exactly what happened. We didn't drop our prices. We cut the timeline. In the right hands, the output is insane. A 3-month project can now take 3 weeks. Designed, built and shipped. And it works. Companies still pay for great design. They just won't wait 3-6 months for it anymore. Upskilling with AI is the best thing any designer or an agency can do right now.

Post attachment
Chris Halaska avatar
Chris Halaska@chalaska·Nov 16

It gives you the most straightforward pattern, which is fine as a starting point. But it's hardly ever the right one for your product, because it's the average of every dashboard it's seen. The last 20% is where I spend most of my time now.

Chris Halaska avatar
Chris Halaska@chalaska·Nov 16

Ask AI to design a dashboard and you'll get four cards in a row across the top. Every time.

Chris Halaska avatar
Chris Halaska@chalaska·Nov 16

AI made speed cheap. Taste is what's short now, and nobody wants to trade the speed to get it. One founder put it to me as wanting someone who can balance "moving fast while having the taste." That's the whole job for me. AI for the speed, 20 years of shipping products for the taste.

Chris Halaska avatar
⟳ Chris Halaska retweeted
Chris Halaska@chalaska·Nov 13

Most designs get lost between the Figma file and the build. Engineers end up guessing the rest. So I design the product and take it into code. It's how I'm building YouSpot's product design system. For funded teams with strong engineers and no design lead: halaska.com/design-engineeri…

0:36
Chris Halaska avatar
Chris Halaska@chalaska·Nov 13

Congrats on $100M ARR 🎉 Proud to have contributed to this; we redid their design system and worked across a suite of new/existing features.

Retell just crossed $100M in revenue run rate. $10M to $100M in 17 months. Profitable. $4.6M raised. From day one, we’ve focused on solving the hardest parts of deploying voice AI in the product itself, so customers can build, manage, and improve their own agents without depending on us. Our bet: the future of the contact center belongs to the teams who run it. Proud of our team. Grateful to every customer who believed in us.

1:47