Andrew Wilkinson (@StartupsILike) header
Andrew Wilkinson (@StartupsILike) profile picture

Investor and Techie. Currently investing for Discover Ventures

New York, NY
Joined August 2017
Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)Photo by Andrew Wilkinson (@StartupsILike)
Andrew Wilkinson retweeted
Here's a Boring Compounder for the last 10 years: $3.87B to $8.20B selling nuts and bolts. $FAST Fastenal, 2015 through 2025: 2015 revenue: $3.87B 2025 revenue: $8.20B, up 8.67% Q1 2025 gross margin: 45.1% FASTBin and FASTVend installations: 129,996, up 12.4% Digital footprint: 61% of sales Customer sites spending $10K+ monthly: 80.7% of sales Q1 2025 dividends paid: $246.7M Fastenal put vending machines inside customer factories and stocked them with the specific parts that factory burns through. Once the machine is bolted to a wall and wired into a customer's purchasing system, changing suppliers stops being a phone call and becomes a project. High-value sites now drive over 80% of sales, up from 78.9% a year earlier.
1
1
1
119
Andrew Wilkinson retweeted
David Barnett was a tenured philosophy professor who glued two clothing buttons to the back of his phone. He had a PhD from NYU and taught philosophy of language and philosophy of mind at the University of Colorado Boulder. In 2010 his earbud cord kept knotting in his pocket, so he drove to a fabric store, bought a couple of large buttons, and stuck them on his iPhone 3 so he could wrap the wire around them. It worked, it looked ridiculous, and his friends would not let it go. Within weeks he was working on a version able to collapse flat, and it took roughly 60 rounds and hundreds of prototypes before he landed on the accordion mechanism. His 2012 Kickstarter asked for $12,000, raised about $18,000, and he burned through it in a few weeks. Then his house burned down in a wildfire, and instead of replacing what he had lost he put much of the insurance payout into making more grips. The first PopSocket sold in 2014, and revenue grew 17,423% over the next three years to reach $169 million. He has since sold over 290 million units.
1
1
2
108
Andrew Wilkinson retweeted
Steve Ells opened Chipotle to raise money for the restaurant he actually wanted to build. He had a diploma from the Culinary Institute of America and two years as a sous chef at Stars in San Francisco under Jeremiah Tower. On his nights off he ate at the taquerias in the Mission District, where a burrito came wrapped in foil and cost almost nothing. The plan was to run something like that in Denver just long enough to pay his father back and bankroll a proper fine dining place. His father lent him about $85,000, and the first Chipotle opened in July 1993 near the University of Denver. The two of them had worked out that 107 burritos a day would cover costs. The opening was rough, partly because there was no menu on the wall and nobody could work out what to order. Once the build-your-own idea caught on, the store was moving over 1,000 burritos a day inside a month, and he never opened the fine dining restaurant. McDonald's took a stake in 1998, then sold out for $1.5 billion after the 2006 IPO. That stake would be worth tens of billions today.
1
1
97
Andrew Wilkinson retweeted
Dave Thomas dropped out of school at 15 and moved into a YMCA so he could work full time. He had been adopted as an infant, lost his adoptive mother at five, and spent his childhood moving from town to town with his father. At the Hobby House restaurant in Fort Wayne he met Colonel Harland Sanders, who became the biggest influence on his working life. In 1962 his boss offered him four Kentucky Fried Chicken stores in Columbus that were bleeding money, in exchange for a stake if he could fix them, and Sanders himself told him not to take it. Thomas found the trouble in the menu, which ran to about 100 items and confused everybody who walked in. He cut it to chicken and a few sides, traded buckets of chicken to radio stations for advertising, and came up with the rotating red bucket sign now bolted to KFC roofs everywhere. By 1968 he sold his share back to the company for $1.5 million, which made him a millionaire at 35. He opened the first Wendy's in Columbus in November 1969 and named it after his eight year old daughter Melinda Lou. He went back to school at 61 and earned his GED, and the graduating class voted him most likely to succeed.
1
2
6
276
Andrew Wilkinson retweeted
4,074 US restaurants is the scale Chipotle is trying to watch for food safety risk from one place. The company confirmed it's piloting a Food Safety Risk Management Platform built on Palantir Foundry. The system pulls together health department inspection scores and pest incident reports. Employee illness reports feed in as well, and each restaurant ends up with a risk score. The goal is to spot trends early and prioritize follow-up before a problem turns into an outbreak. Laurie Schalow, Chipotle's chief corporate affairs and food safety officer, confirmed the pilot. The company hasn't said how many locations are involved, when it started, or whether a chain-wide rollout is planned. The history explains the investment. Chipotle's 2015 E. coli outbreaks hit sales and took the brand years to recover from. It also pulled jalapeños from restaurants in Minnesota over a salmonella concern tied to a supplier. Palantir is better known for government and defense work, and the pilot has drawn questions about how employee health data gets used. Food safety at 4,000+ locations is turning into a data problem, and Chipotle is treating it like one. $PLTR
2
1
3
238
Andrew Wilkinson retweeted
Palantir $PLTR is building Chipotle a system to score food safety risk at each of it restaurants
4,074 US restaurants is the scale Chipotle is trying to watch for food safety risk from one place. The company confirmed it's piloting a Food Safety Risk Management Platform built on Palantir Foundry. The system pulls together health department inspection scores and pest incident reports. Employee illness reports feed in as well, and each restaurant ends up with a risk score. The goal is to spot trends early and prioritize follow-up before a problem turns into an outbreak. Laurie Schalow, Chipotle's chief corporate affairs and food safety officer, confirmed the pilot. The company hasn't said how many locations are involved, when it started, or whether a chain-wide rollout is planned. The history explains the investment. Chipotle's 2015 E. coli outbreaks hit sales and took the brand years to recover from. It also pulled jalapeños from restaurants in Minnesota over a salmonella concern tied to a supplier. Palantir is better known for government and defense work, and the pilot has drawn questions about how employee health data gets used. Food safety at 4,000+ locations is turning into a data problem, and Chipotle is treating it like one. $PLTR
2
2
160
Andrew Wilkinson retweeted
Steven Bartlett went on PBD and explained how he is running his podcast like a data company. 2 things he built: 1. Agent Editor, an AI processing 9 camera feeds and producing final cuts faster than human editors. He tested it for 4 months before trusting it. 2. Guest Radar, a tool predicting how a guest will perform based on their appearances on other podcasts. Some names come back with engagement multipliers as high as 3.1x average.
Steven Bartlett built an AI team that can edit an entire podcast in just 20 minutes and even predict how well a guest will perform. @StevenBartlett is turning podcast production into an AI-powered machine.
1
2
1
5
489
Andrew Wilkinson retweeted
60 Hudson Street has been moving internet traffic for decades from the middle of Tribeca. 32 Avenue of the Americas sits 500 yards away, a full city block of telecom gear behind a landmarked Art Deco facade. Neither building has ever been a local controversy. People walk past both every day with no idea what is humming inside. Now look at the fights over new builds in small towns. The complaints are about water use in dry counties and strain on local grids. Real issues, and specific to the site. The obvious pushback: a Manhattan carrier hotel pulls a fraction of the power an AI training campus needs. Different animal, different scale. Still, the pattern holds. These buildings sat inside the densest city in America for decades without a protest, because the power and fiber were already there. Siting is doing more work in these fights than the technology is.
1
2
2
126
Andrew Wilkinson retweeted
Herb Kelleher co-founded Southwest Airlines in 1967 based on a business plan sketched on a cocktail napkin at a bar in San Antonio. The plan was simple: fly short routes between Texas cities at prices low enough ordinary people could afford to fly instead of drive. Three established airlines immediately sued to block Southwest from flying at all. The legal battle lasted 4 years. Southwest finally flew its first flight in 1971 and the established carriers kept fighting to ground it. Kelleher fought back every time, personally arguing cases in court. Southwest survived and turned the low-cost model into a permanent disruption of commercial aviation. He ran the airline on a philosophy most executives would never attempt: hire for personality, train for skill, and treat employees like the most important people in the business because they are. Southwest became the most profitable airline in American history, recording profit every year for 47 consecutive years before the pandemic. He dressed as Elvis at company parties. He once settled a business dispute with a rival through an arm wrestling match. He ran the most consistently profitable airline in American history by refusing to take himself too seriously while taking the business completely seriously.
1
1
2
108
Andrew Wilkinson retweeted
Todd Graves got the lowest grade in his class for a business plan now doing billions in sales. He was a student at LSU and the plan described a restaurant selling nothing but chicken fingers, fries, Texas toast, slaw, and one sauce. His professor called the menu too narrow to survive and graded it last, and every bank in Baton Rouge agreed. So he went and earned the money instead, working 90 hour weeks as a boilermaker in a California oil refinery. When the refinery work ran out, a coworker told him he could make more fishing in Alaska. He flew up, camped on the tundra for a month waiting for a spot on a boat, then worked 20 hour days pulling sockeye salmon out of Bristol Bay. He came home in 1995 with cash, used it to land an SBA loan, and opened the first Raising Cane's near campus in August 1996. Opening night ran until 3:30 in the morning because the line never stopped, and the menu has barely changed since. Over 900 locations now, and Graves still owns more than 90% of it.
1
1
3
144
Andrew Wilkinson retweeted
Steven Yang left a job at Google to sell replacement laptop batteries on Amazon. He was a Peking University computer science graduate writing search algorithms in California, and in 2011 he quit and moved back to China. To everyone around him it looked like a step down, since cheap accessories sold on a marketplace is not a prestigious business. What he was actually doing was reading the one star reviews. He treated Amazon's complaint section as a product specification, building the version customers were asking for and pricing it under the brand names. Chargers, cables and power banks all came out of the same loop, and his mother put in part of the starting capital. In 2013 they shipped a charging chip called PowerIQ, letting one port work out what device was plugged into it, and that became the reason people bought Anker. Apple dropping the charger from the iPhone box in 2020 did not hurt either. By 2021 Anker was doing over $1 billion on Amazon alone, and it now runs its own brands including Soundcore and Eufy.
1
1
6
1,017
Andrew Wilkinson retweeted
Dane and Travis Boersma had to sell the family dairy herd, so they bought an espresso machine instead. They were third generation dairy farmers in Grants Pass, Oregon, and new environmental regulations made the farm unworkable in the early 1990s. Dane was 38 and had been running a Dairy Queen on the side, while Travis was 21 and looking for anything. Travis suggested espresso, which almost nobody in rural southern Oregon was drinking at the time. They spent about $12,000 on a double head machine and one pushcart, then tested a hundred pounds of beans in the empty milk house. For a month they handed out free samples to anyone who would take one. On February 12, 1992 they rolled the cart out by the railroad tracks downtown, and by 1994 they had four more carts and a drive-through. Dane died of ALS in 2009, when the company was running about 135 stands and doing $50 million a year. Travis kept building it, and there are over 1,000 Dutch Bros locations now.
1
1
1
133
Andrew Wilkinson retweeted
In 1995, 78% of high school seniors went out with friends more than once a week. By the early 2020s the number was in the 50s. The line barely moved for the 30 years before the drop. Then it fell close to 20 points inside a single decade. The timing is the interesting part. The window lines up almost exactly with smartphones going from a gadget to a default. I keep thinking about what happens when those teenagers become the customer. A lot of consumer businesses still assume people leave the house to do things together.
1
2
231
Andrew Wilkinson retweeted
Sandisk 18 months ago: a boring NAND company nobody wanted. Sandisk now: $8.9B in a single quarter. Revenue sat between $1.5B and $1.9B for 8 straight quarters. Then it went vertical: Q4 2025: $2.3B Q1 2026: $3.0B Q2 2026: $6.0B Q3 2026: $8.9B 1Y revenue CAGR of 371.6%. $SNDK
2
1
3
298
Andrew Wilkinson retweeted
Meta added $40 a share in one session. Close: $653.69, up 6.55% The catalyst: Muse, a chat-based personal AI agent launched Monday, built on models developed under Chief AI Officer Alexandr Wang. Next earnings: Oct 28 $META
1
2
266
Andrew Wilkinson retweeted
3 years into his career at Apple, what set John Ternus apart wasn't a big project. It was a screw. He'd joined in 2001 from a small VR startup, after building a mechanical feeding arm for quadriplegics as his senior project at Penn. He started on the Apple Cinema Display. On a supplier visit, he caught a component machined with 35 grooves when it should have had 25. By 2005 he was leading engineering on the G5 iMacs. When his first boss retired and offered him a private office, he turned it down to stay with his team. He became VP of hardware engineering in 2013, then SVP in 2021, running the teams behind iPhone, iPad, Mac, Apple Watch, AirPods, and Vision Pro. On September 1, 2026, after 25 years at Apple, he became CEO, taking over from Tim Cook. A lot of people outside the company had never heard his name until he was named the successor. His advice to Penn grads: walk in assuming you're as smart as anyone in the room, but never assume you know as much.
2
9
1,140
Andrew Wilkinson retweeted
Here's what recession did for this company: $9.5B to $11.8B while the economy was falling apart. Dollar General, 2007 through 2009: FY2007 sales: $9.5B FY2008 sales: $10.5B, up 10.1% FY2009 sales: $11.8B, up 12.8% Same-store sales up 9.0% in 2008 8,362 stores at the end of 2008 450 more opened the year after KKR took the company private in July 2007, a few weeks before credit markets seized. Worst possible timing for a leveraged buyout. It ended up one of the only 2007 mega-deals in the black. Their greatest advantage is the margin. Gross profit ran 30.8% of sales in early 2009 against 25.8% in fiscal 2006. Growth was not bought with discounts. Shoppers came down from Walmart and Target on their own. Cheap goods sold close to home is a model with a hedge built into it. $DG
1
1
122
Andrew Wilkinson retweeted
Melanie Perkins learned to kitesurf because the investor she needed only took meetings at kitesurf retreats. She was 19 and teaching design software at a university in Perth, watching students burn a whole semester learning where the buttons were in Adobe. At 22 she and Cliff Obrecht built a yearbook design tool out of her mother's living room, and it made real money. Then came three years of pitching Silicon Valley on design software for people who are not designers, and more than 100 investors turned her down. She was young, from the wrong country, dating her cofounder, and walking into a market Adobe had owned for two decades. Then she heard an investor named Bill Tai was coming to Perth to kitesurf, and she found out he ran his networking events at kitesurf retreats. So she learned to kitesurf. Tai never invested himself, but he brought her into his network, and the seed round came out of it. Canva now has over 265 million monthly users and $4 billion in annual recurring revenue. Perkins and Obrecht still hold about 18% between them, and they have pledged to give most of it away.
1
4
497
Andrew Wilkinson retweeted
Apple's first foldable iPhone debuts today, expected to start at around $2,000 John Ternus @johnternus runs his first launch event 8 days into the CEO job. $AAPL
1
6
250
Andrew Wilkinson retweeted
Larry Page and Sergey Brin were Stanford PhD students in 1996 trying to solve a specific academic problem: how do you rank the importance of web pages relative to each other. The answer they came up with was to treat links between pages like academic citations. A page linked to by many important pages was itself important. They called the algorithm PageRank. It worked better than any search engine available. They built a search engine called BackRub, renamed it Google, incorporated in a friend's garage in 1998 with $100,000 from Sun Microsystems co-founder Andy Bechtolsheim, and started. They tried to sell Google to Yahoo for $1 million in 1998. Yahoo passed. They tried to sell it to Excite for $750,000. Excite passed. They kept building. Google went public in 2004 at a $23 billion valuation. Google is now part of Alphabet, worth over $2 trillion. Yahoo is a footnote. Excite does not exist. They started with an academic idea about how to rank pages and accidentally built one of the most valuable companies in human history by solving the problem better than everyone who had tried before them.
1
2
309