## Wednesday, October 22, 2008

Since Google Reader went away I lost the easy way to share interesting articles. This post is a place to drop comments with links to interesting articles and things that don't rise to the level of a full post.

1. Interesting article about applying group testing to pandemic surveillance. I've thought it would be neat to apply to verification testing of complex software.

2. The City is not Neutral
"Cities are complex adaptive systems—or more accurately, many systems of systems that allow the unique civil, social, and economic conditions that originally drove cities’ development and continue to allow the city to survive as an optimized hub of a large civilization. Like other complex systems, when it is touched, it changes..."

3. So grateful that I've had awesome, meaningful jobs my whole career. Sad to hear so many folks think their work is BS.
On the Phenomenon of Bullshit Jobs: A Work Rant

4. How to Make Your Arguments Stronger (Hint: Longer Is Not the Answer)
dilution effect: "...when you introduce irrelevant or even weak arguments, those weak arguments reduce the weight of your overall argument."

5. Most Research in Deep Learning is a Total Waste of Time - Jeremy Howard | AI Podcast Clips

“you get just a whole lot of research which is minor advances in stuff that’s been very highly studied, and has no significant practical impact”

6. Online edition of the The Feynman Lectures on Physics

7. Youtube playlist for The Mechanical Universe videos

8. National Committee for Fluid Mechanics Films

There's a playlist for all 39 on youtube.

I especially like the ones on vorticity.

1. this is how good investment decisions should be evaluated in any era: "executives should link investment and capture more fundamentally and consider the gain in probability-of-win and probability-of-go that each marginal investment dollar can bring"

it doesn't make sense to invest in either something that's a sure win or a sure loss, evaluating on the margin (where can the marginal dollar have the biggest impact on pushing win probability towards 1) is where the smart money should already be

9. The Dictatorship of Data< Robert McNamara epitomizes the hyper-rational executive led astray by numbers.

1. "one standard deviation closer to God"

10. "If you can't make up numbers that make it look like you could be successful, why should they believe anything else you said?" chuckle, chuckle
Ernestine Fu: All You Need to Know About Venture Capital

11. Determining How Much Testing is Enough: An Exploration of Progress in the Department of Defense Test and Evaluation Community

1. Mackinder, H.J., The geographical pivot of history, The Geographical Journal, Vol. 170, No. 4, December 2004, pp. 298 – 321

12. Venturing through the Pandemic
"Lockheed Martin Ventures lays claim to the earliest investments, dating back to 2007, though 36 of the 56 companies Lockheed Martin Ventures has invested in have joined the portfolio just since 2016.

Airbus Ventures started investing in 2016 and lists 39 companies in its portfolio, including the space debris-mapping company LeoLabs, which operates radars in Alaska, New Zealand and Texas to track objects in low-Earth orbit.

Boeing established HorizonX Ventures the following year and has invested in 34 startups, including nine in 2020.

The companies usually don’t disclose specific investment amounts for competitive reasons. HorizonX says typical investments in earlier-stage companies range from $1 million to$10 million.

The big three may make investments that are less about a near-term profit, as the executives tell it, and more about advancing technology that could further certain broader strategic goals, either for the corporation itself or for the startups in its portfolio."

13. The Weakness of Modern Monetary Theory
There are theoretical and empirical limits.

1. "...what makes MMT unique is that it's the only school of thought (should it be even considered as such) to combine calls for not only increased, but unlimited deficit spending with the argument that central banks can print money to pay off these debts without prompting inflation. There is simply no truth to this assertion. In fact, the reality is much the opposite: As interest costs consume increasing portions of revenue, at a certain point, the government cannot afford to borrow any longer. If the government resorts to printing money to pay its debts, inflation will ultimately follow.

In searching for radical theories to advocate radical policies, MMT apologists have inadvertently come to endorse an approach to economic theory that not only flies in the face of decades of economic research and historical precedent, but would be devastating if ever tested. By tempting progressive politicians into endorsing "dangerous" and "obviously indefensible" economic theories, MMT advocates are setting the stage for potentially disastrous policymaking."

1. "The AI competition is also a values competition. China’s domestic use of AI is a chilling precedent for anyone around the world who cherishes individual liberty. Its employment of AI as a tool of repression and surveillance—at home and, increasingly, abroad—is a powerful counterpoint to how we believe AI should be used."