Wharton Professor. Teach, research, and practice innovation, entrepreneurship, and design.
Philadelphia
Joined June 2011
- Tweets837
- Following403
- Followers4.6K
- Likes563
Nicely designed clip, right? Integral spring and hinge. Nice blend of function and beauty. It’s from 350 BCE! ~2400 years old. (A fibula, for holding cloak closed.) Fascinating exhibit of Celtic craft @harvartmuseums through August 2.
Karl T. Ulrich retweeted
What would you do, as an industrial designer, if your client asked you to optimize a tool—and you couldn't? That was the challenge faced by ID firm Goodwin Hartshorn, when… core77.com/posts/143645/Mini… #IndustrialDesign #KitchenTools #ProductDesign #DesignInnovation #JosephJoseph
I ran into one of my heroes at the breakfast buffet! Donald Knuth made algorithms into a rigorous science and wrote the canonical treatise on it, and invented TeX along the way. DEK remains whip smart at 88, and very much engaged with AI developments. He was attending Gathering 4 Gardner (wizards, math, magic, and more).
Karl T. Ulrich retweeted
Check out this great new paper by @AngeloCarlino3, @AliciaWongel, @duanlei3008, Edgar Virgüez, Steve Davis, @energymorgan, and @KenCaldeira here: sciencedirect.com/science/ar…
Karl T. Ulrich retweeted
Canada’s new government just released the Housing Design Catalogue.
It has 50 standardised housing designs that builders will use to streamline approvals, cut through red tape, and get millions more homes built, faster.
How many grams of protein are harvested per animal life? Dairy generates 390,000 g per life. That same amount of protein requires 800 chickens. Tuna may surprise you. Baby octopus is probably the worst thing on the menu. My latest paper detailing this analysis via link in reply.
reply guy FAQ1: yeah, but is a chicken life the same as that of a cow? [Of course not. See Section 2.1]
reply guy FAQ2: yeah, but what about animal welfare? [Good question. See Section 4.3]
reply guy FAQ3: yeah, but why not go vegan? [Good question. You do you. For those who choose animal protein, my analysis may be informative.]
Karl T. Ulrich retweeted
Link: Octopuses Fall for the Rubber Hand Illusion stevestewartwilliams.com/p/o…
Wonder how duties get charged and collected? We @makerstockinfo just paid our first new tariffs. Here is the invoice from our freight agent. Freight from Shanghai to NY is USD 6900 (up 50% because of whipsawing Q2 demand). Overland freight ("drayage") is normal $1250. Tariffs $12,739, exactly 50% of what we paid the factory for the goods themselves. This cost goes into cost of goods, which we of course account for in prices. The importer pays the agent for everything to get the goods from the foreign port to the US destination. The agent then pays the actual shipping company and the government.
Karl T. Ulrich retweeted
"hyper precise prompts to describe what you want", is absolutely the BEST strategy.
🔥 Many YCombinator AI startups prompts are super detailed (e.g. 6+ page prompts) with XML tags and meta-prompting techniques.
e.g. Parahelp's customer support agent prompt is 6+ pages, meticulously outlining instructions for managing tool calls.
---
⚙️ Key Learning from this doc
→ Top AI startups use "manager-style" hyper-specific prompts—6+ pages detailing task, role, and constraints. These aren't quick hacks; they’re structured like onboarding docs for new hires.
→ Role prompting anchors the LLM’s tone and behavior. Clear persona = better alignment with task. Example: telling the LLM it's a customer support manager calibrates its output expectations.
→ Defining a task and laying out a plan helps break complex workflows into predictable steps. LLMs handle reasoning better when guided through each sub-task explicitly.
→ Structuring output using markdown or XML-style tags improves consistency. Parahelp, for instance, uses tags like <manager_verify> to enforce response format.
→ Meta-prompting means using LLMs to refine your own prompts. Feed it your prompt, outputs, and ask it to debug or improve—LLMs self-optimize well if given context.
→ Few-shot prompting with real examples boosts accuracy. Startups like Jazzberry feed challenging bug examples to shape LLM behavior.
→ Prompt folding lets one prompt trigger generation of deeper, more specific prompts. Helps manage workflows in multi-step AI agents.
→ Escape hatches instruct LLMs to admit uncertainty. Prevents hallucination and improves trust.
→ Thinking traces (model reasoning logs) and debug info expose the model’s internal logic. Essential for troubleshooting and iteration.
→ Evals (prompt test cases) are more valuable than prompts themselves. They help benchmark prompt reliability across edge cases.
→ Use big models for prompt crafting, then distill for production on smaller, cheaper models. Matches quality with efficiency.
Karl T. Ulrich retweeted
I have so many thoughts on this, as someone who has run large EPD orgs.
One of the toughest parts about doing an AI operating transformation inside a prod/eng/design org is that people start operating out of their lane and it stresses SO MANY PEOPLE out.
PMs can vibe code prototypes -> eng is stuck with aspiration + code that aren't production ready
Designers can add in more polish and craft -> scope starts to creep
Engineers can vibe-code new features -> who asked you to build that
But the reality is: this isn't really an AI problem. The inefficiencies of of "stay in your lane," "this is too complicated," "i own the prioritization" have always existed, they've just been codified into the operating manual of most orgs so they feel right ("it's how things are done") and they make everyone on the team feel happy, because it maximizes individual loci of control (PM owns this, design decides that, you get whatever eng decides to build) vs the velocity of the system.
AI starts to crack the facade of these fiefdoms, in more obvious and frequent ways, and you start to experience dumb conversations like below often, with a stronger sense that there's a better way to build.
But humans cling to their old ways! And the operating model you've built is calcified in such a way, it's easier to give up and stay in your lane vs. push for a new model.
But I have faith. I think if you can change your *culture* you can change how you build, for the better.
Repeat after me:
- make experiments cheap: try that fancy design coded w AI, you may get further than you think. or you can throw it out without too much waste
- get good with AI tools: you can marry a vibe-coded demo with a production-ready codebase + AI IDE and actually get pretty far with quality code
- there are no lanes: welcome people coding, designing, PMing even if it's not their official title
- make it fun: your roadmap is not life or death. the more people are having fun, vibing (literally), and don't feel the crushing weight of an organization-level anxiety disorder, the faster they can go
Finally, you can just build things!
(And be nice to your colleagues, please.)
21 RMB for Peking Duck delivered to my hotel in Dalian, China. That’s 3 USD nominal and 5.78 USD PPP. Meituan platform. How possible? High volume for the shop, pre-made, delivered to automated locker (in hotel lobby), and yes, wages under 30 RMB/hr. Lunch for 6 next day from KFC (not my choice!) w/ espresso drinks 300 RMB (82 USD PPP). Order the duck!
Seen on the beach today in Gold Coast (Aus). Dredger Trud R replacing some of the 4 million cubic meters of sand removed by Cyclone Alfred. 5 megawatt peak power using 1200 liters/hr diesel. No worries, mate.
Most goods imported from China (unless you are Tim Cook) are now 55+% not 10%. The legacy tariffs stack. An item I care about (a specialty plywood): historical tariff 3.2%, Section 301 tariff (Trump 2018) 25%, Fentanyl tariff 20%, reciprocal tariff 10% = 58.3%. There is no domestic supply 🤷♂️
