How a tiny TV console made me develop a passion for machine learning

In the spirit of the season to be thankful, I would like to share a story with those of you who love video games, starting innovative companies and investing in them.

In the fall of 2015, I landed in the US and after a month long search for a place to live, I ended up renting a small studio from Robert Brown. I was pleasantly surprised to find out that my first home in the US was also the home of the person who unknowingly influenced me to study computer science and contributed to me wanting to live in Silicon Valley.

You wonder how is that possible? I grew up on the Black Sea coast of Romania and one day around the age of 5, my dad and I went to buy groceries, but instead, we came home with an Atari 2600. We were no longer hungry the second we started playing the first game, Frogs and Flies. Both my dad and I were instantaneously hooked, to my mother’s despair, and always competing. The game was on! So it was really easy for me to recognize The Atari Middle Earth Pinball in the kitchen of the house I was visiting here in Los Altos Hills, while looking for my first residence in Silicon Valley. Asking about the pinball machine in the kitchen, I found out that the man standing in front of me was previously the head of the microelectronics division at Atari and directly responsible for our family’s beloved 2600.

‘Wow! This is great! go ahead develop it!’

At the age of 14, Dr. Brown had a friend who was working at a TV store and he started working there too, repairing TV sets back in 1951. For my millenial peers, in those days TV sets were black and white, and the biggest one was 16 inches and had tubes, imagine that! After graduating from Stanford, he worked for North American Rockwell in pattern recognition - in particular speech recognition - for 5 years. He then joined a few friends from graduate school to work for Fairchild Semiconductor doing automated layout of integrated circuits. After that he got a call from former coworkers at North American Rockwell who were starting a new company. He joined them but things didn’t work out as planned and the team was acquired by Standard Microsystems, developing integrated circuits. He did simulation, testing and programming. Then it happened that a friend of his from the startup went to work for Atari, and he came to Robert asking him to consult and do a test program for what he was developing at Atari at the time. These were integrated circuits for coin operated machines. In the process of that engagement Robert came up with an idea: He observed a guy who was feeding two channels of a stereo system into a TV set, and some really interesting patterns formed. Robert quickly noticed the problem was having to sell the whole system which was a big console TV with the stereo built in it and so forth. So he thought what if you could just do a device that hooks to your TV at home and generated the patterns instead. Then he got to think ‘hey they’re doing video games’, how about doing a device that lets you play the video games on your home TV. So he went to Atari with that idea and when he showed the video music prototype to Joe Keenan, Atari’s president. Joe might have been a little high in that particular moment and exclaimed ‘Wow! This is great! go ahead develop it!’. Atari was 2 years old when Robert joined full time. And that’s how Robert started working on a first product which was the Pong game that came out in ‘75.

A real relaxing company with a lot of freedom

Back in those days, companies required their male employees to wear a white shirt, tie and a sport coat. Atari was a loose company compared to those companies. Robert describes it as a real relaxing company with a lot of freedom. And sometimes they would be drinking beer or smoking some pot but was only at brainstorming sessions and never during work. As one might have guessed, all the technical employees were males. I asked Robert particularly how it would have been for me to work at Atari, as a female engineer, and he admitted that it was a male chauvinistic culture, and sometimes they would name projects internally after girls they liked, such as Arlette. I am glad that the software industry has evolved and is getting better and better and I want to give a shout out to all of you who are accelerating this progression. The struggle is real! Robert in his own way helped to pave the way when he hired a female programmer but does not remember any technical person at Atari who was a female before then. He had 32 engineers working for him as he headed the microelectronics division at Atari.

Steve Jobs is crazy

Robert recalls Steve Jobs was there at the time doing consulting work, and was hired by Al Alcorn, the chief engineer. Nolan, Atari’s founder, wanted to reduce the number of integrated circuits in the coin operated game so he told Steve he would give him $2000 for every integrated circuit that he could delete from the system design. And so Steve went to Woz and said ‘hey I talked with Atari and they said for every integrated circuit we can design out, he will give us $500 and I’ll split it with you’. Woz designs out 5 or 6 integrated circuits but Atari could not use it because it was not a straight flow through and was very hard to debug so they decided it was not practical to use. Robert remembers when Steve decided to leave Atari and start his own computer company that Al Alcorn said he’s crazy. And boy was he wrong! Steve told Robert at the time, that he was afraid that Atari would crush Apple after Atari launched their own 400 and 800 home computers. But Atari built their home computer as a closed system and that’s how it died, against Robert’s recommendation of making it open so people could develop software for it. Steve then shared with Robert that he should not have been concerned at all about Atari’s home computer.

Sell it before you release it

Before Atari released the Pong game, they were trying to sell it to someone else in the industry. They went to Mattel but Mattel was not interested. They went around to other stores and they got the same answer until they tried the TV department at Sears. Sears also said no, and so did the stationery department but the guys did not give up and said ‘it’s a tennis game, why don’t we go to the sporting goods department?’ and the person there really liked it. The device they showed him was ten times bigger and the person was a little skeptical they could be able to put all that electronics into a little chip, but they convinced him. That’s how Sears became the first customer. Atari sold a Pong game with the Sears brand name and one with the Atari brand name. Once the product came out that Christmas in ‘75, there were people lined up at the Sears store trying to get on a waiting list to get the game. They were planning to make 50 thousand units but ended up making 250 thousand. Nobody else had it, it was THE thing.

The game designers were the real unicorns

Atari knew they couldn’t keep doing dedicated games, and needed programmable game cartridges. At that time there was only one developer assigned per game, and that person had to love games, had to be a good game designer, know what made a game really addictive, and had to be an excellent programmer because the cartridges had 4k of memory and they would spend 80% of the time trying to reduce the code size to fit into 4k of memory. That was the real issue, trying to fit the game in the cartridge, but they would get it done, and with no degradation of the game itself. It would take about 9 months to make a really good game.

One thing Robert thought he did well was to find great people. He recalls one of his employees saying “he knows how to make himself look good, hiring these people” and Robert took it as kind of a negative remark but he thought about it later and thinks he should have said thank you because that was his job. His job was to hire people who made him look good, not for the purpose of making him look good, but if he hired good people that was the result. And that’s why he was there, to interview and hire good people.

The old game formula that powers instant game hits even today

Many companies were trying to develop games at the time, including Quaker Oats, and many were terrible because they had not realized the formula for a great game: you want something that anyone can walk up to and start playing immediately, you want something that people can get better and better at, and that when they get better and better the rewards get bigger and bigger. A classic example is Breakout, Robert’s favorite game, in which you have to smash two walls of the playfield, one brick at a time. Today he thinks the games are very complicated and the games that become instant hits nowadays are those that follow the old formula. He is by the way still trying out new games even today and shared his experience with a passion that is hard not to notice. The kind that manifests itself when people light up talking about what they love, that kind of passion that inspires.

The rug merchant and the beginning of the end - replacing the CEO is not always a great idea

Atari attracted people because if they loved games, so this is where they had to be. It was as simple as that. Things changed when Warner bought the company and decided they were going to install their own CEO and brought in Ray Kassar, the head of Burlington Industries, a textile company. You wonder what textile has to do with video games, and so did Robert and his peers with regards to ‘the rug merchant’ as he was referred to at the time. He thought ‘my god, he is totally non technical and he does not understand what is all about’. Moreover, he had a bad idea that he could fire and hire the game programmers as if they were a dime a dozen. And so out of the 32 engineers Robert had hired, only 2 were left, and those were test engineers and not even design engineers. This was a year after Robert left Atari, he was pushed out once the rug merchant came in. Robert recalls they were also promised a big bonus before Warner acquired the company, 25 cents for every game sold to be distributed to the engineering group. But when Warner took over, which was when Atari just broke even, they refused to give the bonus even if they told Robert to mention it when they’d hire people. And when Robert confronted them they said they did not remember and played dumb. After the engineers left, they were trying to hire consultants to make games and they were paying them big money, however that did not translate to success. He recalls they decided to launch an ET game because ET was so popular. The consultants developed the game in 6 weeks. And the game was so bad, that they eventually had recalled all game cartridges and took them to the dumpster and poured concrete over them.

Market always wins

Atari was so popular that 15 companies including Quaker Oats (lol) started to make games for the 2600 console. But after a while, the market got saturated and the video game business came to a crashing halt and that was the end of Atari. Robert was caught up in that with Starpath, a startup he had founded after leaving Atari. Starpath were able to add more RAM to the system in the carthridge and therefore have twice the vertical resolution than anyone else could have by writing self modifying code. And although their games were ranked number one by reviewers in game magazines, Starpath failed because the market for games crashed. 

Market always wins, again

Starpath started in ‘81 and got funding in ‘82. When Robert started raising, the market for games was doing very well. While raising money for his own company, Robert noticed that the VCs were focused more on the team than on the product. Mayfield Fund invested in them because they had missed investing in the game market at that time and it was a no brainer to invest in Starpath: they could do bigger games, they had the experience from Atari and they were incredible game developers. Starpath launched 11 or 12 games and the secret of their short lived success were again the game engineers. The games were not tested with end users, the programmer was the end user, they wanted to do a game better than anyone else. They were the maker and the target market, they were the whole package, game designer, programmer and user. Several of the engineers at Starpaths were junior college dropouts 17, 18 old who loved to code and play games. At the time when Starpath was around there were game magazines and they had reviewers and Starpath had the best reviews because of the higher resolution and the game play. But then the market for games crashed. What they had not anticipated was that the market had become saturated. The market had suddenly met demand. More product was dumped on the market and it was not going anywhere in ‘82 and ‘83. Robert recalls the VCs pushing a guy on him to become the CEO of Starpath. A guy who had had one success, but sometimes success only happens because you’re in the right place at the right time. There is a lot of luck involved and it does not mean the same person starting another company will have the same result. The new CEO did not do much, he would take naps during the planning sessions and come back again “refreshed”. Regardless of this issue, Starpath still failed because of the market. Starpath was the number one in reviews and had an explosive growth, and yet was around for only for 2 years.

I wanted to share Robert’s story as a way of thanking him for positively influencing my life and the lives of many others. Passion is the common element for success, be it passion for repairing TV sets, passion for developing the best goddamn game or hiring the best design engineers. And bringing in a BigCo CEO who has no passion for the startup might not be the best choice. Are you passionate about your work?

His story also taught me that one should think about the market development even when the market is hot right now.

Thank you Robert for being the best landlord ever and letting me play with all those Atari games at home and for hiring that female programmer at Atari and paving the way for us! You rock!