Are you using something with a modern microprocessor on International Women’s Day? (If you’re not, but somehow able to see this post, talk to a doctor. Or a psychic.) You should thank Dr. Lynn Conway, professor emerita of electrical engineering and computer science at Michigan and member of the National Academy of Engineering, who is responsible for two major innovations that are ubiquitous in modern computing. She is most famous for the Mead-Conway revolution, as she developed the “design rules” that are used in Very-Large-Scale Integration architecture, the scheme that basically underlies all modern computer chips. Conway’s rules standardized chip design, making the process faster, easier, and more reliable, and perhaps most significant to broader society, easy to scale down, which is why we are now surrounded by computers.
She is less known for her work on dynamic instruction scheduling (DIS). DIS lets a computer program operate out of order, so that later parts of code that do not depend on results of earlier parts can start running instead of letting the whole program stall until certain operations finish. This lets programs run faster and also be more efficient with processor and memory resources. Conway was less known for this work for years because she presented as a man when she began work at IBM. When Conway began her public transition to a woman in 1968, she was fired because the transition was seen as potentially “disruptive” to the work environment. After leaving IBM and completing her transition, Conway lived in “stealth”, which prevented her from publicly taking credit for her work there until the 2000s, when she decided to reach out to someone studying the company’s work on “superscalar” computers in the 60s.
Since coming out, Dr. Conway has been an advocate for trans rights, in science and in society. As a scientist herself, Dr. Conway is very interested in how trans people and the development of gender identity are represented in research. In 2007, she co-authored a paper showing that mental health experts seemed to be dramatically underestimating the number of trans people in the US based just on studies of transition surgeries alone. In 2013 and 2014, Conway worked to make the IEEE’s Code of Ethics inclusive of gender identity and expression.
A good short biography of Dr. Conway can be found here. Or read her writings on her website.
If you’re seeing this on any kind of computing device on International Women’s Day, you should thank Dr. Grace Hopper, rear admiral of the US Navy. Hopper created the first compiler, which allowed for computer programming to be done in code that could more closely resemble human language instead of the essentially numerical instructions that work at the level of the hardware.
These “higher level” languages are what are typically used to create all the various programs and apps we use everyday. What have you done today? Word processing? Photo editing? Anything beyond math was considered outside the domain of computers when Hopper started work.
An article about the “most ridiculous startup ideas that became successful” has been making the rounds on social media. It amused me, mainly because the “ridiculous” ideas used to summarize each company are more like strained ex post facto descriptions that describe what they currently do, not the starting business model.
- Facebook was not meant to be another Myspace. It started as a way for college students to communicate with each other (after a very brief life as a “hot or not” thing for Harvard dorms). If you’re a Millenial, ask your parents if they ever looked at Classmates.com. Odds are that they have. Myspace was a public site where 13-year-olds made 90s-esque web profiles that were open to everyone, including 40-year-old men pretending to be 13-year-olds. That Facebook did not require this degree of openness has been part of its success.
- Dropbox seemed like the first major file transfer program I heard of aside from Google Docs. As this XKCD shows, we’re still desperately working on file transfer and so almost any idea could go. (My current solution is Google Drive)
- Amazon took off a lot after eBay drew people online. Amazon started around the time of the dot com bubble, so it’s not like investors needed much rationalization before investing in websites. But if you think about it, the basic starting idea kind of makes sense: Amazon could get virtually any book for a customer without wasting money on inventory costs. Also, Amazon hasn’t turned a profit in years because it tries to keep expanding, so maybe we should be wary of calling it a success for now.
- Virgin was founded not long after the airline industry was deregulated, so the timing isn’t crazy.
- I know virtually nothing about Mint or Palantir, but the idea of a company being really dependent on defense contracts is actually not uncommon.
- Craigslist is a classifieds web site in a time when newspaper classifieds are slowly dying. Investing in it seems really reasonable. And actually, it doesn’t seem to be pulling in a lot of venture capital money. The one major outside investor is eBay.
- iOS isn’t even a company or a standalone product. Why is it on this list?
- The whole point of Google was that its indexing algorithm was almost completely different than other search engines at the time. Does the author not remember how bad search results were in the 90s? Also, Google grew out of Larry Page’s dissertation, so it’s not like pitching was done before the algorithm existed.
- Part of PayPal’s appeal is that it’s more secure to give just one website your financial information and use that for purchases than to give your credit card information to a new person every time you make buy something online
- LinkedIn totally confused me in college, but now I appreciate separating my professional and social networking activities. And evidently lots of companies do use LinkedIn for recruiting since they can sort-of target appropriate people better than random Internet ads.
- Tesla actually does work with other car companies on some models and does have a goal of providing electric car equipment to other manufacturers to help mainstream electric cars. And considering that it was founded in 2003, its existence predates the cleantech “backlash”.
- 2/3 of SpaceX is owned by Elon Musk. And SpaceX doesn’t just want to be a commercial NASA (and even if the author finds this really weird, I would invite him to read almost any science fiction talking about space colonization). It plans to do commercial satellite launches as well, which are big business now.
- Firefox is the work of a free software group. Which is mostly funded by a non-profit (and a company that makes money, but that reinvests nearly all of that into the non-profit).
- Honestly, the only crazy ideas here seem to be Instagram and Twitter. And people still seem unsure of how those are supposed to make money so maybe we’ll find that their current structures are ridiculous.
Of course, the reason this list is so popular is because people seem to love counterintuitive ideas proving some experts or conventional wisdom wrong. It’s like Malcolm Gladwell applied to entrepreneurship. And just as wrong.
So Microsoft has just announced a pretty awesome speech recognition/speech translation program. More on this later, but the video is really cool. If you want to skip to the translation bit, jump ahead to 7 minutes in and you’ll see English to Chinese text, and then later it goes to Chinese audio (I assume it’s Mandarin?).
There are a few obvious limitations – You can definitely tell he is speaking slower than normal. A few times he gets really excited and you can see the English speech recognition accuracy really drop. And even speaking slowly, it’s not perfect. But it’s pretty good overall (I can’t comment on the translation, since I know no Chinese). Also, they say his voice was used for the Chinese audio, and I believe it, but it doesn’t sound incredibly “unique”. To me, it just sounds like it’s about his pitch, but that’s it.
Update: So what makes this different than other translators? You may have noticed he described previous work as using “hidden Markov modeling” and this is a “deep neural network”. Markov chains are basically networks of probabilities. For instance, we can use a Markov chain to describe your lunch behavior if you’re really ritualistic. Let’s say you and I are co-workers. We’re in separate wings and are good acquaintances, but maybe not super close friends, so if we run into each other, we’ll eat together, but otherwise we won’t. There’s a 60% chance our morning meetings end such that we run into each other right before lunch. If you eat by yourself, there’s a 30% chance you grab pre-made sushi from the cafeteria and eat at your desk, a 50% chance you eat something from the grill in the cafeteria, a 10% chance you go out and get barbecue for lunch and finally a 10% chance you go to the Mexican restaurant. If you meet me, there’s a 10% chance we eat take-out sushi from the cafeteria, a 30% chance you we go to the cafeteria grill, a 20% chance we go get barbecue , and a 40% chance we go to the Mexican place.
Someone who knew all of this could figure out how often you’re likely to eat at each thing. The equivalent of a hidden Markov model in this case might be a co-worker who always see your leftovers at your desk and then tries to work backwards to figure out the probabilities of these events happening. As the above XKCD points out, sometimes the most probable thing following doesn’t always happen. And that’s the big limit to a Markov model. You can only work with the most recent state.