The Future of Information Work: Strategic Planning

Grand Challenges for engineering - What do you think?

Sun, 17 Feb 2008 18:00:36 GMT

The National Academy of Engineering has launched its open innovation project to start thinking about how to address the new Grand Challenges that a select committee believe are crucial to the century ahead.

Here are the Grand Challenges for engineering as determined by a committee of the National Academy of Engineering:

They are taking open comments on these ideas. Here was my first post:

I think we need a more distributed, organic method that integrates public policy makers, citizens and the organizations that co-exist in a given area. Rather than pushing the burden to over burdened government, we need bring together stakeholders in a way that may drive this to a very very local level - not just streets, but community association, even individuals. Yes, there are grand needs for large improvement, but even those can be more effectively handled in an organic way by seeking emergent planning and construction methods.
Think about putting a city's infrastructure into a computer, modeling the current needs and then projecting, using scenario planning, along multiple different futures, and then have the system evolve, using genetic algorithms or some other process, a plausible plan within the constraints. In this way, science can help depoliticize the planning process and providing transparency into how various constraints: natural, technological and political, were handled in the various models. The system could even look at financial capabilities and search for funding models from available sources, and while also suggesting a timeline.
This does beg the issues of many science fiction stories in which a computer plans the future and people simply accept its dominion over them. I am suggesting here, that software is capable of being an effective tool in negotiating human issues often fraught by imperfect information. Not only does the software provide emergent views of the future, but it can provide the forum for negotiation and cooperation – perhaps brokering an auction among interested parties for paying for various portions of a project – those who choose to live on an island end up disproportionately paying for a bridge, but may have little to do with a rural road hundreds of miles from them. The model can also look at things like wealth distribution and wear, so that perhaps the rural road, given less traffic, doesn’t get addressed as often unless there is a natural disaster.
If all the power lies with the political infrastructure, they physical infrastructure will always suffer. We need to find ways to bring rational alternatives to the table that go well beyond competitive bidding by several interested contractors who are never as apolitical as they may seem. And then of course, we will need to write the software to look for bias in software models, which isn’t on the list of big challenges, but may be one of the next ones.

I encourage you to get involved. I don't know where this will go, but they have raised important issues, and, as I mention in my post, they may have missed a few implications. They need to ask the question: if we solve this, what new challenges does the solution create? An important question to any quest like this.

A futurists work is never done :)

Don't Play Games with the Future- Bueno de Mesquita and the Myth of the Good Forecast

Thu, 15 Nov 2007 03:31:38 GMT

How do we predict the future. Well, according to the cover story of Good magazine the answer is game theory as advocated by Bruce Bueno deMesquita. My first impression, and my second, is read The Black Swan. Bueno de Mesquita is building a myth. He has accidentally predicted the future based on some well chosen variables. He is building models, and some of those models are pretty sophisticated. I like game theory. It is a wonderful input, but it is not correct because as sophisticated as the models are, they are sophisticated enough to model all the complex variables associated with a multi-national (or even personal) event. If you really engineer well, you can get a workable model that may be right, but can you generalize the model, predict other things with the same model. No, you need to understand you subject, to build in relationships and feedback loops. You need to model the environment and therein lies the problem.

We don’t have enough CPU to model everything. That is our great future myth, that with computers, we can eventually build a model that predicts behaviors. We are wonderful prophets of the past now. We can tell exactly what people who do yesterday, but any number of things can change behaviors tomorrows – new technologies, new policies, new incentives. As a sales person what they sell when a company changes an incentive plan to pay them more for selling one thing over the other. What do they sell. Could we predict the behavior without the policy change? Probably not.

That doesn’t mean that Bueno de Mesquita shouldn’t be building models. I do worry about when I see the US Military and the CIA in the article. I do hope that they realize this is one input. I’m sure they have read the Black Swan, but the power of myth, of stories, of expertise, is sexy and appealing – especially when it comes with data that substantiates claims. I just hope the buyers of the myth aren’t running anything I’m involved in when a model goes south, and it will. They better have tested some alternative futures through scenarios and other techniques because betting on a sure thing is very risk in today’s world.

How to Tell if a Forecast Is Credible (or how to make a credible forecast)

Wed, 05 Sep 2007 12:47:18 GMT

After my lengthy commentary on the BusinessWeek Future of Work series (see their articles here: The Future of Work) I think it is only fair that I put down my expectations for an article that purports to forecast the future.

Tell people what you don’t know. This creates a great list of items for multiple outcomes and it really make it look like you did your homework. Pretty simple: here are the areas where prediction leaves use with near complete uncertainty, and here is a range of possible outcomes.
Don’t just write the article. Too hard to find the forecast. Have multiple industry pundits and other knowledgeable people comment on how things might turn out, and then use the those fancy print capabilities to visually map the possibilities for people. Be sure to include, in the commentary, the trigger events that people think will cause their forecast to happen (or not to happen)
Disclose the framework for the forecasts. A long list of articles with no connective tissue is useless for decision making. If everybody isn’t using the same framework (say a shared set of scenarios) then the predictive power, and decision making ability of the forecasts again, are nearly useless.
Highlight disagreement. On some topics (and I know print media is limited compared to the web) print opposing points of view. Give people information about the reasons that people believe what they do and make it clear that the consensus is or isn’t there on a particular forecast.
For polling, don’t ask people what they think will happen. Interesting, but not that useful. Rather, design the survey so that the answers indicate certain characteristics about the future that may be different than the questions themselves. Remember, questioning people about the future from a finite list of choices means the survey has systematically eliminated innovation, because only the things they thought about are in the survey. Also, most surveys presuppose a single future, and therefore do not provide any insight about alternative futures based on trigger events that would shift the entire context (open global markets vs. a sustained economic cold war, vs. a real war, vs. some regional mix – answers to questions would be different if the future was different. For instance, the survey said most people were worried about China. In a particular future, the quality issues facing China may lead to a decline in their economic strength. If that future is presented as an alternative, the survey results are likely to be very different than the current commonly held belief that China will be the economic superpower of the 21^st Century.

So if you are forecasting, or reading a forecast, take this as a high level list of things to look for. If you don’t see them, then take the forecast with a grain of salt (all forecasts should be taken with a grain of salt, but if you follow this list, at least the underlying faults in the forecast will be transparent.)