The Dressler Blog

When you build digital stuff all day, you develop opinions. Lots of opinions.


Digital Trends

On the Suckiness of Passwords Passwords suck. But they are only a symptom of a deeper problem, which is how we store and share our personal information. Currently, our personal information is everywhere. We hand it over to company after company in distinct accounts. Each of these accounts are ostensibly protected by a password. However, as we learn time and again, companies are actually terrible at protecting their customer’s information and terrible at notifying customers when their information has been exposed to hackers. Internet security specialists are starting to wonder if the blockchain – the anonymous, distributed ledger that powers cryptocurrencies like Bitcoin and Ether – might be a better way for users to manage their personal information. Already, the state of Illinois is looking for ways for a unique identifier to replace the current (badly broken) social security identification system. Rather than hand your information over to a company, you could allow them temporary access through the blockchain. This is technology that exists to enable transactions in a trustless environment and consumers are increasingly untrusting of the security of large companies. Why does this matter? It being the holiday season, I find myself blithely inputting my credit card number into companies’ websites with depressing frequency. Any one of these systems (built by fallible human beings) could be hacked and my identity could be auctioned off in some former Soviet republic for the price of a case of Vodka. Why do we agree to such an idiotic system? Credit cards might have made sense in an era where fraud was strongly localized. But if my identity was stolen tomorrow, I would have no idea which site had coughed up my details and I would have no idea how to work with the police in Murmansk to fix the problem. Companies have proven themselves crappy (and deeply conflicted) stewards of their customer’s security. It’s time to rip and replace the current system. In a nutshell: Bitcoin offers more security than Uber . Read More Next Generation Software Development My eldest daughter (all of six years old) is deeply interested in robots, unicorns, and fairy princesses. I make up for my ignorance in the latter two categories by answering her questions about robots. The other day on our walk to school, she asked me how someone could program a robot to do things. “Well,” I explained with fatherly smugness, “right now, a software developer would have to create instructions for everything the robot did. Detailed instructions that would include every tiny movement and every single variable, thousands of lines of code long. But that won’t be the way it is done for that much longer.” “Why?” she asked sweetly. I continued my explanation: “There’s a new kind of software development that depends on neural networks. Neural networks power a technology called machine learning. Sometimes it’s also called artificial intelligence. In the future, rather than writing code for every little thing the robot does, developers will use neural networks. They will just establish general goals, what they want the robot to do and let the neural network figure out the most efficient way to get it done.” “Why are they called neural networks?” Me (beaming with barely-concealed pride): “Great question, sweetie! They are called neural networks because they are networks of distinct processing units that are built to resemble the neurons in our brains. Neurons are the things that allow us to think and some people believe that neural networks allow machines to think. Do you have any other questions?” “Yes, Daddy. When a unicorn shoots magic out of its horn, is the magic in the horn or is it all around us and they just use their horns to focus it?” Why does this matter? It doesn’t take a software developer to know that my answer is actually a gross oversimplification of a complex issue. Namely, that neural networks are changing the way software is developed. After last week’s blog post about Population-Based Training (PBT) of neural networks, one of my friends sent me a link to a blog post by Andrej Karpathy, who is the Director of Artificial Intelligence at Tesla. Karpathy believes we are on the verge of what he calls Software 2.0, an era in which software will be built altogether differently due to the advantages of neural networks. Karpathy’s analysis (unlike mine) is detailed, specific, and informed by his years of experience developing complex software for some of the most advanced technology companies on earth. If you are interested in understanding how neural networks will transform programming, I suggest you read him. If you are interested in understanding unicorn magic, I am willing to forward any questions to my daughter. n a nutshell: Neural networks will make software more powerful, but also more opaque to developers since they will address tasks on their own without human instruction. Read More The Seed Slump Like a lot of people kicking around the technology industry in New York, I spend a lot of time having meetings in the coffee shops around Union Square, Flatiron and up into the new neighborhood called NoMad. It’s inevitable that you overhear a lot of conversations. Entrepreneurs, as a group, are high on confidence and low on volume-control. For the last year, these overheard discussions have taken on a slightly-hysterical edge. The word was out. Early stage funding was drying up. Sure, there were a couple crazy funding stories out there – Masayoshi Son alone is good for one or two stories a quarter. But the small investments that most startups depend on to build up their MVP and prove a market just weren’t that easy to come by. So I wasn’t surprised when Fred Wilson, the dean of east coast venture capital wrote a recent post in his hugely influential blog AVC that explained why we were in an early stage investment slump. According to Wilson, 2012-2016 represented something of a bubble in early stage financing as numerous investors flooded the market with easy money. However, that bubble led to an undervaluation of early seed investment. The valuations were unfavorable to the investors and the hit rate was far too low to justify further investment. The good early stage investors saw better valuations and possibilities in Series A investment, so they moved up-market. The result is that early stage investment has slowed down significantly and entrepreneurs are hustling harder for less money – raising the volume in coffee shops on both coasts as a result. Why does this matter? I hear a lot of silly reasons why the tech industry in the United States creates so much wealth and innovation. Everything from the pioneer spirit to manifest destiny to “something in the water in Silicon Valley.” As a sometime entrepreneur myself and a friend of other entrepreneurs, I can tell you that the sources of all this tech energy are forgiving bankruptcy laws and ready availability of investment capital. Of these two, investment capital is the most important factor. (No one actually plans on bankruptcy.) Experienced technology investors know that you don’t make money in technology by chasing unicorns. This type of investing leads to a lot of “me-too” companies – the Uber of power tools, the Warby Parker of baby carriers, the Netflix of gifs. Real value can be found in those companies that create technology so obscure that you need to do a lot of due diligence to understand if such a market even exists. Experienced technology investors also know what a reasonable valuation for an early stage company should be. If they don’t get their valuation, they stay away. They haven’t been getting their valuations because of a glut of (you’ll have to forgive my bluntness) dumb money in early stage investment. That dumb money is also attracted to the most simplistic, “elevator pitch” style of investment. So naive investors pumped money into easy-to-understand companies and got burned and the bubble finally burst. As Fred Wilson points out, that’s an opportunity for all those experienced technology investors to get back into early stage investment. In a nutshell: You’re going to have a hard time getting funding for your Snapchat for Baby Boomers startup. Read More

Give us your email to sign up for our weekly Dressler Digital Trends. Stop trying to keep up and start getting ahead.