Automation is more entrenched in some industries than in others. The availability of affordable computer hardware and better software means that benefits are being harnessed by more sectors.
The legal profession is the latest that is set to be revolutionized by automation software, with armies of lawyers and their paper-driven systems potentially being replaced by artificial intelligence computing. One eye-opening moment for me was when I saw that a law office had actually printed out every email that I had ever written or received. Including the mail attachments, that was three moving-boxes full of paper.
Imagine: the law office actually had someone read through that! Clever “e-discovery” software could have analyzed these documents in a fraction of the time and at much less cost. The inevitable side effect is that some lawyers and paralegals will lose their job.
Compare that to our field of design automation, where the tedious jobs of printing, checking and hunting for patterns have already been automated for decades. When it comes to the impact of automation on society, EDA is different than other industries in one key respect. What you see generally happening in the real world is that automation is putting people out of a job –– the evidence is there in the growth of self-checkout counters at supermarkets, and the way that banking is now done with far fewer human interactions. These industries follow the pattern where the contribution of automation means that certain jobs disappear.
But in our field, even though our business is to automate the job of the designer, our great friend here is Moore’s Law, which makes chips more complex at a faster rate than you can automate problems out of it. The net effect is that more designers are needed with each generation of chips.
We have many more difficult and complex techniques needed to squeeze everything out of the chip, especially in terms of power usage or battery power. That requires EDA companies to get better and better at implementing and quickly trying a variety of combinations.
EDA’s progress can barely keep up with the relentless speed of Moore’s Law. The impact of electronic design automation is, therefore, simply necessary and not in the least detrimental. Without the automation, we would not be able to design these hugely complex chips, at least not at an acceptable cost.
We do a few things well in EDA. First, we are used to dealing with massive amounts of data, and we know how to manipulate it. In other fields, such as the legal profession, this pattern-matching process is much less developed.
Some of the techniques that we have innovated and used in EDA are applicable to other fields. A good example is in the field of biology. The Janelia Farm Research Campus is a pioneering research center in Ashburn, Va., and part of the Howard Hughes Medical Institute for biomedical research. The research undertaken at the Janelia lab seeks to understand the brain in all its aspects. Researchers at the lab are now using derivative of EDA tools to reverse-engineer the brain of the fly.
To put this work in perspective, the simplest organism currently being examined in biomedical studies is a small worm that has only 302 neurons. Even so, the workings of its brain are not entirely understood. As the next ambitious step, biomedical engineers at Janelia are seeking to understand the brain of the fruit fly, which has 100,000 neurons.
The Janelia researchers are using EDA techniques for extracting chip layouts to reverse-engineer the network circuitry of flies’ brains. You need a lot of EDA tools to understand patterns at this level of complexity.
Human beings have a whopping 100 billion neurons. That is six orders of magnitude more complexity. It will almost certainly take beyond our lifetime to fully understand the human brain.
It’s great to know that EDA is now helping this fundamental research into how the brain works. It’s just one of the ways in which EDA’s skill at understanding complexity can be enormously valuable in other fields. And our field of automation can create jobs, not eliminate them.