6 May 2015

Goldman's Self-Defeating Case Against a Programmer

By Leonid Bershidsky

Here's to Sergey Aleynikov, the former Goldman Sachs programmer convicted Friday of stealing the bank's high-frequency trading code, for accepting no compromises and persisting with a legal battle that still could put him in prison. By having the U.S. court system recycle the shaky six-year-old case, he is showing others like him the absurdity of working for old Wall Street banks in an era when a tiny startup, or even a guy trading from his home, can run circles around them.

Lawyers can have fun arguing about the legal aspects of Aleynikov's case. There's plenty to gnaw on: the use of a pre-computer statute on "unlawful use of secret scientific material" to convict him, the spirit and the letter of double-jeopardy law (Aleynikov was cleared of federal charges in the same case before the state of New York pursued him), and whether a judge should dismiss a case when the jury clearly struggles to understand the subject matter. There's the separate matter of whether a company should be allowed to slap its own copyright on open-source code that its programmers barely change, which is what Goldman apparently did with the code Aleynikov downloaded on one of his last days at the bank.

But all that has to do with law, not finance or programming -- the two areas for which Aleynikov's case is especially important. 

Imagine you're a highly skilled programmer, perhaps an immigrant like Aleynikov, unfamiliar with the intricacies of the U.S. legal system and not 100 percent confident in your English skills. You're probably more at ease dealing with machines and the abstract tasks they solve when interpreting code than with people -- otherwise you'd do something else for a living. 

Here's how the coverage of Aleynikov's trials reads to such a programmer: A big bank offers a decent salary to a coder. When he takes the job, he's confronted with a legacy system so bloated, it can't really compete with more nimble upstarts in the same business. (Goldman Sachs wasn't among the top high-frequency trading firms in 2009, and it still isn't. The top ones aren't household names.) He is encouraged to use open-source code to achieve results faster, but he's not allowed to give anything back to the community that developed it. His grander ideas are shelved, because the bank just wants quick fixes to its old, messy platform. When someone offers him a better job, free of the legacy system's constraints -- one that pays four times as much, as Aleynikov's new job with Teza Technologies did -- the bank enlists the FBI's help in going after him for saving some code on an outside server. Then, after the coder spends a year in jail and the case against him collapses on appeal, the bank keeps going after him, using its considerable political clout.

This is hardly a legal interpretation of what happened, but it tells a programmer that joining an established Wall Street bank is a bad idea. It's better to sell one's skills to a startup built by refugees from such banks. Many HFT firms started that way, because their founders hated the banks' slow, rigid systems. Such Wall Street escapees are more inclined to trust and cooperate with their programmers.

Aleynikov's case resonates beyond the small HFT world to the finance industry as a whole, which is undergoing a technological revolution akin to the one that's reshaped the news media. This revolution is being carried out by programmers, and they need employers who understand the nature of their work and how likely they are to be headhunted and asked to reproduce their best work elsewhere -- a frequent occurrence in tech. 

Traditional banks tend not to be such smart employers. That, and the fact that startups begin with a clean slate rather than oodles of legacy code, accounts for the increasing competition that banks face from tech-based outsiders -- including high-frequency trading outfits, robotized asset management firms, app-based retail banks, money transfer companies and payment systems. 

Global investment in financial technology ventures tripled to $12.2 billion last year. That's not much money by Wall Street standards, but startups can do much more with it than traditional banks can, because it's often easier to build from the ground up and they have no need for armies of lawyers and managers.Banks that don't want to end up obsolete, like print newspapers, need to be friendlier to tech professionals. They need to listen to them when they say legacy systems need to be ditched, not patched. And banks should definitely avoid trying to put programmers in jail for doing what they always do. Goldman may be hoping Aleynikov's case will set an example to others like him, teach them to respect the bank's rules. Instead, it will tell good programmers to go elsewhere and perhaps, sooner than necessary, make Goldman obsolete.

No comments: