I think the lack of engineers in the US government is becoming a bigger liability every year. Here science is changing the globe at an ever increasing pace and our congresspeople are baffled and confused by the word "encryption."
There aren't many reasons an engineer would want to jump into the political fray besides altruism. It's like choosing to join the largest, most corrupt, bureaucratic, change-averse, and valueless company that ever existed. For lawyers, that probably sounds like a dream come true.
My perspective as an engineer: Yes, there are huge important problems I would like to help solve. No, I don't think I'm willing or capable of solving that class of problems in a system like that.
- Terrible consultants, middlemen, and lifers making progress too hard or impossible.
- Constantly changing or unclear requirements
- Making penny-wise, pound-foolish decisions for political reasons or budgeting idiosyncracies
- Resistance from lawmakers since it would give more government influence to people from states that lean towards some of other party (put more bluntly, Republicans don't want to give a bunch of liberal or left-leaning people from California, Washington, and New York more influence)
- The government only seeing the value in surveillance/military software that people are likely to refuse to work on. Not sponsoring software that actually directly helps people domestically
What I think we need instead is something like the US Digital Service, along with some other government programs, rolled into a Department of Technology that will hopefully have a completely different culture, payscale, etc. compared to the rest of the government. And maybe even something like jury duty, but for software people, engineers, scientists, etc. so it's more representative of the general public and not the old-school government culture.
Can you elaborate on this? I hear lots of criticism around inept and ill-intentioned consultants and “govvies” alike, but I think a lot of it is a bit unfairly harsh. There are certainly some adverse incentives (chiefly that it’s easy enough to coast along with minimal effort in these jobs if you’re inclined), but I’d say the slight majority of individuals in this space are qualified, capable, and motivated to improve the world via good work—technical or otherwise. Slowly but surely, projects get done and positive change is made.
> Not sponsoring software that actually directly helps people domestically
This really isn’t true. Every US agency has technology initiatives, some more effective than others. See the work that the Department of Veterans Affairs does, as one example. The US government also spends ~$750B on grants annually, much of which goes to fundamental research as well as tech projects. If you really want to dig in, there’s an API to access data on every item of spending (e.g. contract awards, grants): https://www.usaspending.gov/#/
I can mostly only echo hearsay, however I have been tangentially involved in government-focused projects. The big issue I see with the government getting involved in tech is that they don't know what they really want, they don't know how to filter out bullshit/upsells, and they'll frequently overpay for things they don't need and underpay for things that would be very helpful.
I'm not one of those people who wants to defund the government or privatize everything, however I do think the government incentive structure needs to be redesigned and become less cushy while simultaneously becoming more lucrative, accompanied by a huge culture change. The government isn't inherently inefficient, it's only like that because we let it be, one has only to look at other countries (see: Estonia, China) to see how much more effective we could be here.
Regarding actual spending, I want to measure that by its effectiveness, not by the total amount spent. The government could blow all its money on tech initiatives and EMR upgrades and have nothing to show for it, and I wouldn't call that helping people domestically. To address your examples, a lot of those grants are for defense contractors (which only indirectly helps people, at best, and is itself a pretty corrupt old-boys industry). Others are for things like direct academic research though, which I do think is decently well run in general, but isn't really a direct government relationship to tech.
These are good points (as were your original ones). Thanks for the thoughtful reply—I’m just trying to push back slightly on the “government/contractors are all incompetent” narrative where I see it, since I don’t think that’s the root of the problem (the DMV area is the most highly education region of the US, as one counterpoint). Discussion around incentive structures is a lot more productive.
I didn’t give very specific examples with grants, but I just wanted to draw attention to that as an often overlooked and major source of government spending. The federal grants process does have multiple stages of project performance monitoring/evaluation, though that process itself could be made more efficient. Also, sometimes grants do directly support software projects with public benefits. Just as a single example, Stanford DeepDive (now Snorkel) is a system for knowledge base construction that has been used to fight human trafficking, among other things: http://deepdive.stanford.edu/showcase/apps
Do we need AI engineers passing laws or training courses for politicians?
Can the AI be treated like a black box [0] and the regulation still be effective?
I think in some cases the answer would be yes. A simple thought experiment demonstrates this.
Example: Consider regulation for AI in relation to job losses. Does one need to know the details of AI tech to develop regulation effectively? I would argue not.
The economic impacts can be analysed. At some tipping point, job losses Lead to a cycle of:
reduced demand -> reduced Revenue in companies -> reductions in jobs -> reduced demand and so on and on.
To regulate this, one would have to model the economic impacts as best as one could. Understand what jobs are most likely to be lost, in which industries. Understand the key things to monitor (I.e replacement job rate. Assuming jobs are created in other industries). And pass regulation to mitigate some of the risks or make sure the number of Job losses never hit the tipping point. None of these items leads me to believe the person passing the regulation requires knowledge of AI at an engineers level.
There is the Conant-Ashby Theorem or good regulator theory [1]. This states “Every good regulator of the system must be a model of the system” [2]
In this instance the regulation is to deal with ECONOMIC impacts. Therefore an economic model is needed, while impacted by the tech the tech does not need to be modelled. Thus my conclusion.
This poses an interesting question, could all eventualities be catered for in the regulation? There is other theories that could help us answer this question.
Happy to hear different or contradictory views.
[0] https://en.m.wikipedia.org/wiki/Black_box#Other_theories
You don't need to know the ins and outs of the mathematics to understand what encryption does, but you should be able to reason about how data is encrypted and what is necessary to decrypt, and what law provides you to request decryption. You sure as hell should also be able to grasp that, like the atomic bomb, once out in the world it's there to stay and you can't really prevent it from being used.
As for economic impact - I find that measure does not really get ahead of the problem. In the case of machine learning if you're regulating economic impact you're decades late. The first-movers already have the data sets, the training has occurred and now you're living in their world. I think it's important to understand a bit more upstream so that you can get ahead of these kinds of issues.
Furthermore, a baseline understanding of these algorithms and their potential biases / impacts of misclassification is important to understand to ensure we don't compromise society by their mis-use, or shift scales in a terrible way unintentionally. The earlier the better, and economic impacts are probably a lagging indicator, or even worse considered a huge boon for those areas that have companies doing this work but using labor from elsewhere to train their models.
Lets play some word-games:
* Doctors : Disease, Ebola, Health Care costs
* Accountants: Corporate Taxes, Tax Loopholes, Economic Stats
* Programmers: Encryption, Websites, "Internet is Tubes", Infrastructure as a Service
* Army Generals: War against ISIS. War against Taliban. War against (any other enemy here).
* Civil Engineers: Bridge design and safety, Urban Smart Growth, Gentrification, Water Rights management across rivers.
* Energy Engineers: Power Plant construction, Nuclear Safety, Solar vs Coal investments.
* Farmers: Patent law applied to GMOs and ownership of genetics, Legal Insecticides and whether or not they are damaging the local ecology, legal limits to water-consumption across various states and locals
--------
Believe it or not: US Politics involve more than just programmer issues. We care about programmer issues because Hacker News attracts programmers. But as soon as we leave the subject of programming, we (as a community) will become blubbering idiots in the face of Hospital issues, Child-care laws, foreign policy and Tax-fraud.
The entire point of a Congressman is to specialize in the ability to talk to a wide variety of experts. Congressmen are "experts at politics": the ability to make a decision WITHOUT being an expert in a subject.
-----------
Our job, as experts in our field, is to form a cohesive argument that Congresspeople understand. Well, that's the point of lobbying: hiring specialists who know how to talk to Congressmen, so that our interests are communicated more effectively.
And lets be frank: there are plenty of lobbyists who are actually representing technology's point of view at the moment. Google, Facebook, Microsoft, IBM, and other big-tech all ensure that Congress is keenly aware of the issues affecting our field.
Other fields are not quite as lucky as ours. I don't think we have any right to really complain about the Programmer's role in modern US politics.
All joking aside, I think there is a certain way of thinking that comes from a technical education that could be used more in government.
The Chinese government seems a lot more rational, and it's not just because they have authoritarian control. Our government could be far more rational too.
But that's not the important part of government. The important part of government is:
1. Identifying parties who might be affected by a decision.
2. Informing / talking with said parties.
3. Deciding if said parties have conflicts of interests / biases to consider.
4. Making a decision afterwards.
Expertise in a particular subject is... kind of not needed. There are always more experts out there. The only "expertise" needed is again: the politics. Being able to identify conflicts of interest and account for them.
Sure, a CEO doesn't need to be an engineer because the CEO doesn't do any engineering, but he understands the fundamental forces that shape the success of the engineering projects and he respects the process.
Like Steve Jobs?
> CEO with an engineering background
I mean, Lisa Su and Jensen Huang are great. But I'd argue that Elon Musk is terrible as a CEO.
I mean, you're asking the very basic Bill Gates (Programmer background) vs Steve Jobs (non-programmer) question. Obviously there's a difference, but I wouldn't necessarily say that the lack of engineering background hurt Steve Jobs.
-------
I'm not entirely sure if an engineering background is necessary to be a good or bad CEO. The most important thing is being able to lead your workers. For some, it means being a technical leader, understanding the code and product.
For others, like Steve Jobs, it means being a visionary and making the right business choices. Another great non-technical CEO would be Warren Buffet, a great CEO and visionary: being able to predict the future of the American economy and place investments in the right place.
There's lots of ways to lead and inspire others. We programmers worship technical leaders, because that's our strength. But I'm aware of many other styles of leadership, and respect that too (Steve Jobs, Warren Buffet)
Congressmen are experts at getting people to vote for them. The problem is that the people who step up to the challenge of winning this popularity contest aren't necessarily most qualified to be calling the shots.
I'd like to see more STEM workers in general running for office. They'd certainly pique my interest more than yet another lawyer.