Trevor Blackwell on Viaweb, Robots, and Early Y Combinator

概览

Jessica Livingston and Carolyn Levy interview Trevor Blackwell, Y Combinator co-founder and longtime friend of Paul Graham and Robert Morris. The episode traces Blackwell’s path from Saskatoon to Harvard, then into Viaweb, Yahoo, robotics, and the early days of YC.

A central thread is Blackwell’s builder mindset: he was drawn to web-based software because it allowed rapid updates, built hardware like the Viaweb “bogometer,” kept servers alive through improvised infrastructure, and later pursued dynamically balancing walking robots because science fiction had promised them and reality had not delivered.

The conversation also revisits early YC from Blackwell’s perspective: the first Summer Founders Program, printed applications, long interviews, the first Mountain View dinners, the robots that became part of YC lore, and the improvised nature of the original office space.

分段落总结

[00:24] Introducing Trevor Blackwell

[事实] Jessica introduces Trevor Blackwell as her friend and a Y Combinator co-founder. [事实] The hosts say they want to clarify details mentioned in the earlier Paul Graham interview. [事实] The conversation begins with Blackwell’s path from rural Canada to Harvard’s PhD program.

[01:08] From Saskatoon to Harvard

[事实] Blackwell grew up in Saskatoon, Saskatchewan, where only a small number of people around him were interested in computers. [事实] His mother got him an Apple II, and he read about Jobs, Wozniak, and Silicon Valley. [事实] He studied engineering and computer science at Carleton University in Ottawa. [事实] Work with Bell Northern Research connected him to H.T. Kung, who became his Harvard advisor.

[03:01] Catching Up at Harvard

[事实] Blackwell says arriving at Harvard was an eye opener because many people around him knew more than he did. [事实] He and Robert Morris had the same advisor and nearby offices. [事实] Paul Graham asked Morris who the best programmer in the program was, and Morris pointed to Blackwell. [事实] Blackwell says he was strong at practical programming but initially insecure about theory.

[04:34] Joining Viaweb

[事实] Paul Graham introduced Blackwell to Viaweb’s idea: helping businesses sell things on the internet. [事实] Blackwell believed catalog companies would eventually need to be online, which was a bold idea in 1994. [事实] His own experience ordering electronics parts through paper catalog forms made online commerce seem obviously useful. [事实] What excited him most was writing server-side web software instead of boxed Windows software.

[06:12] Why Web Software Felt Revolutionary

[事实] Blackwell describes pre-internet software as requiring CDs, boxes, version numbers, and long release cycles. [事实] He disliked the idea that bugs might take a year to fix because software was physically shipped. [事实] Viaweb’s model let the team run software on their own Unix systems and update it whenever they wanted. [推测] This was one reason Blackwell saw Viaweb not only as an ecommerce company but as an early example of software as a service.

[08:13] The Demo That Pulled Him In

[事实] Graham showed Blackwell a browser-based demo where clicking links changed data on the server. [事实] Blackwell thought that someday all software might work that way. [事实] On his birthday, he built his own version in Smalltalk to test the idea. [事实] After showing it to Graham, he began working on Viaweb more seriously, though formal stock and salary arrangements came later.

[09:58] Finishing a PhD While Selling a Company

[事实] Blackwell finished his PhD, but Viaweb and his dissertation overlapped in a stressful way. [事实] He graduated on the same day Viaweb signed its acquisition deal with Yahoo. [事实] He left the graduation ceremony in crimson robes to sign the acquisition documents. [事实] His dissertation centered on more robust ways to measure system performance.

[12:31] The Bogometer

[事实] The “bogometer” was a mechanical dial display that showed Viaweb metrics such as hits per second and later revenue. [事实] Blackwell initially thought the idea was bogus because moment-by-moment hits seemed less important than daily trends. [事实] The device became useful because sudden drops could reveal outages or other problems. [事实] It also became a highlight when investors, customers, or potential employees toured the office.

[14:47] Early Robot Building

[事实] Blackwell had been building robots since middle school. [事实] His first robots were clunky, including one made from an oil drum, army surplus motors, switches, and eventually an arm. [事实] The bogometer was not his first mechanical or robotics project. [推测] The episode frames his robotics work as a long-running obsession rather than a career detour.

[16:24] Viaweb’s Server-Room Adventures

[事实] Viaweb’s servers were in Blackwell’s office, which was nicknamed the “hot tub” because it was so hot. [事实] The team stacked multiple window air conditioners to cool the room. [事实] During a Cambridge power outage, Blackwell drove to Costco, bought a gas generator, brought it back, and tried to power the servers. [事实] The generator was loud and impractical, and the team still lost power for a period.

[21:35] Keeping Viaweb Running

[事实] Many Viaweb merchants wanted orders faxed to warehouses, so the company used a growing rack of fax modems. [事实] As usage increased, server uptime became increasingly critical. [事实] Blackwell describes making a dangerous male-to-male power cord to connect running servers to a UPS. [事实] The hosts highlight his familiar line, “Okay, it should work now,” as part of Viaweb lore.

[24:26] Working with Paul Graham

[事实] Blackwell says Graham often insisted there was a right way to do something. [事实] Blackwell would sometimes disagree, try another approach, and later realize Graham had been right. [事实] Jessica reports that Graham described Blackwell as “an extremely powerful engine with a rather small rudder.” [推测] The dynamic suggests a productive but opinionated partnership between strong builders.

[25:50] Moving to Yahoo

[事实] After Yahoo acquired Viaweb, Blackwell moved to Mountain View and helped relocate employees and infrastructure. [事实] He was excited to be in Silicon Valley near companies such as Apple, Hewlett-Packard, and Silicon Graphics. [事实] Yahoo initially felt startup-like and gave the Viaweb team room to operate. [事实] Over time, Yahoo became more like a big company, with slower processes and formal user research.

[29:52] The Ad Product Yahoo Did Not Launch

[事实] Blackwell built a prototype ad-buying product for Yahoo Store merchants. [事实] His idea was that small merchants would buy online ads with credit cards in smaller amounts. [事实] Yahoo’s ad sales organization was focused on large ad buys from major companies and did not embrace the model. [事实] Internal pricing rules forced him to show an $80 CPM instead of the actual market-like $8 CPM, which helped kill the product. [推测] The story illustrates one reason Yahoo missed opportunities that later resembled Google AdWords.

[34:39] Starting Anybots

[事实] After leaving Yahoo, Blackwell started a robot company the following Monday. [事实] He later says he would advise people to take a break between major efforts. [事实] He wanted to build human-sized walking robots because science fiction had promised them and they still did not exist. [事实] He began with tools, aluminum, and prototype legs in his basement before moving into more space.

[37:56] Dynamic Balancing and Robot Design

[事实] Blackwell contrasts Honda’s ASIMO-style walking, which used large stiff feet and careful geometry, with dynamic balancing. [事实] He wanted a robot that could walk more like a person, including on uneven or soft terrain. [事实] He used pneumatic cylinders because they were springy, durable, and closer to muscle-like force behavior. [事实] He wanted hardware that could survive falls, jumps, and impacts without destroying gear teeth.

[42:01] Boston Dynamics and Commercial Reality

[事实] Blackwell says recent Boston Dynamics work is well ahead of where he got. [事实] He says there were periods when he was ahead and periods when Boston Dynamics was ahead. [事实] He did not have a clear commercial use case for the walking robot. [事实] He considered military demand but was reluctant because of pacifist leanings and dislike of bureaucracy.

[44:30] Dexter, Monty, and YC Lore

[事实] Jessica says she first talked extensively with Blackwell at a party because he was showing videos of his robot. [事实] Dexter was the walking robot, while Monty was the upper-torso robot mounted on wheels. [事实] Monty was tele-operated by a person wearing special controls and could manipulate objects. [事实] The robots became part of early YC dinners, giving visitors something memorable to see.

[47:37] The Call to Start Y Combinator

[事实] Graham called Blackwell with the idea of becoming the kind of investors they wished they had when starting Viaweb. [事实] Blackwell recalls how opaque fundraising had been during Viaweb. [事实] The initial plan was to try the Summer Founders Program for one summer. [事实] Blackwell helped read applications and assess whether applicants’ technology could work.

[49:49] The First Applications and Interviews

[事实] The first applications were emailed in and transformed into PDFs for easier reading. [事实] Blackwell wrote software to reformat the applications with clearer question formatting. [事实] The team printed more than 300 applications and scored them manually. [事实] Early interviews were much longer than later YC interviews, but the team learned they often formed opinions quickly.

[52:00] The First Batch

[事实] The hosts note that the first batch eventually included founders associated with Twitch, Reddit, and Sam Altman. [事实] Blackwell says he did not have a preconceived expectation for how many companies would work. [事实] He attended more than half of the first summer’s dinners despite travel from California. [事实] He found the dinners energizing and liked products such as Reddit and Kiko.

[54:05] YC Comes to Mountain View

[事实] After the first summer, Jessica and Graham wanted to run a Silicon Valley program so no one else would become the YC of Silicon Valley. [事实] They used part of Blackwell’s Anybots space in Mountain View. [事实] The building had been a warehouse and office space, with drab interiors, green carpet, low ceilings, and fluorescent lights. [事实] Kate Courteau oversaw renovations, but the city issued a stop-work order before the first dinner.

[56:01] Improvised Office Renovations

[事实] After the stop-work order, some work continued informally. [事实] Blackwell wired many of the lights himself because official electrical work could not proceed. [事实] The first dinner happened while paint was still drying, and people were told not to lean on the walls. [推测] The early Mountain View office embodied YC’s improvised, builder-driven culture before it became institutionally polished.

[58:10] Justin.tv, Early Streaming, and Office Chaos

[事实] Blackwell remembers the Justin.tv team using his lab bench and soldering iron to repair their streaming backpack. [事实] The backpack had multiple cell phone antennas because phone-based streaming was not yet practical. [事实] During one YC dinner, pranksters ordered many pizzas for Justin, and Jessica paid for them. [事实] The office’s limited bathrooms became a problem as YC batches and demo days grew.

[62:03] The Sneaky Bathroom and Demo Day Stories

[事实] The back warehouse bathroom was informally protected for Anybots people and was initially dirty and utilitarian. [事实] Jessica recalls offering that bathroom to Demi Moore during a crowded demo day because the main bathroom lines were too long. [事实] The hosts describe the situation as embarrassing but memorable. [推测] These stories show how early YC’s rapid growth outpaced its physical infrastructure.

[62:50] The Note Card Story

[事实] Blackwell says Paul Graham’s version of the note card story made it sound more eccentric than it was. [事实] Blackwell took inspiration from John McPhee’s process of organizing research on note cards. [事实] He tried using note cards for thesis and class notes but later judged the system terrible. [事实] He denies using note cards as party conversation prompts.

[65:35] Segway Polo and a Home-Built Segway

[事实] Blackwell built his own Segway-like vehicle after the Segway came out because he understood balancing from robotics. [事实] He wrote about it online and was invited to a Silicon Valley Segway polo group that included Steve Wozniak. [事实] He brought his home-built version to games, where it had exposed metal and could turn much faster than a normal Segway. [事实] The group eventually did not let him keep playing with it because of its advantages and risks.

[68:16] What Blackwell Is Working on Now

[事实] Blackwell says he has packed up Anybots and shipped the robots to England. [事实] He is still working on robot software. [事实] His current focus is applying recent machine learning advances to robot body control. [推测] His current work continues the same theme from the episode: combining mechanical intuition, software, and long-term curiosity about autonomous robots.

[70:07] Post-Interview Reflections

[事实] Jessica and Carolyn say Blackwell clarified the note card story. [事实] Jessica says the Segway polo connection helped her get Steve Wozniak for Founders at Work. [事实] The hosts emphasize that Blackwell was fun to work with, technically helpful, and central to YC’s early culture. [事实] Carolyn says the robots were part of YC folklore and that she also received a robot tour when she first visited.

播客点评/总结

This episode is valuable because it gives a first-person account of several early internet and startup milestones from someone who was not only present but building the infrastructure, tools, and prototypes behind the scenes. Blackwell’s stories make Viaweb, Yahoo Store, Anybots, and early YC feel concrete rather than mythic.

The strongest parts are the details: the bogometer, the generator, fax modems, application PDFs, wet paint at the first Mountain View dinner, and robot demos in the back of the room. These anecdotes reveal how much of early startup history depended on improvisation, technical range, and tolerance for messy execution.

[推测] The main limitation is that the conversation is more anecdotal than analytical. It does not deeply evaluate Anybots as a business, the broader robotics market, or the full strategic reasons Yahoo failed to capitalize on some opportunities.

[推测] This episode is especially suited for listeners interested in startup history, early web software, YC’s origins, robotics, and the psychology of unusually capable builders.