In Crichton's 2002 speech which cites the Gell-Mann amnesia effect ( http://larvatus.com/michael-crichton-why-speculate/ ) he argues that:
(1) newspapers and other media sources should not quote experts who predict that the 2002 United States steel tariffs will affect GDP or employment, because no one can predict the future, and you should ignore any predictions you read experts making which are quoted by newspapers. Note that in retrospect ( see https://en.wikipedia.org/wiki/2002_United_States_steel_tarif... and links ) research has showed that the tariffs adversely affected GDP and employment, as some experts predicted.
(2) other predictions which people made are wrong and, with the benefit of hindsight, they should not make them. Note that Crichton cherry-picks only those predictions which turned out to be inaccurate.
In retrospect there is a bit of irony in his being forced to "predict" certain future predictions which would turn out to be wrong. The one specific future prediction he happened to choose happened to be somewhat accurate.
Yes, Crichton does use the latter half of the speech to argue that global climate change is not happening, or if it is happening it's good for humanity, and anyway it's considerably less important to humanity's fate than changes in the Earth's magnetic field strength.
What he does not say in his speech is "I have analyzed predictions made in news articles for the past X years, and judged the accuracy of N predictions -- Y% of them were accurate. Frustratingly none of them expressed any degree of confidence in their predictions so I treated all equally for this analysis. This percentage is [no worse than guessing | worse than guessing, so you should expect the opposite of what is predicted with weak/strong confidence | better than guessing, so you should expect what is predicted with weak/strong confidence]. Here are my data so you can see for yourself."
Did Crichton do his homework? Or did he just give up in dismay and cherry-pick some examples of failed predictions? His argument is rhetorically compelling but it's based on anecdotes rather than data and it's hard to trust.
A very quick Web search shows that with respect to opinion columnists (not the same as newswriters), this work has been done at an undergraduate level -- see https://www.hamilton.edu/news/story/pundits-as-accurate-as-c... , https://www.hamilton.edu/documents/an-analysis-of-the-accura... -- and has found that some specific opinion writers tend to make predictions which are accurate and some do not.