I get nervous when I see articles about GMOs that aren't published in Nature, Science, or their ilk, but I'll bite.
Genetic modification (engineering implies a precise science, it isn't) is useful for several reasons in crops:
1. We can increase hardiness of plant to resist frost and disease more readily,
2. We have the potential (which isn't always realized) to increase the yield and nutritive value by increasing the ploidy,
3. We can control where and in what conditions the crop is capable of growing, preventing spread to undesired locales and taking over natural plant life.
People hear genetic engineering and immediately think science fiction, but the truth of the matter is that genetic manipulation of plants only really involves three things:
1. Adding resistance genes already present in other organisms. This includes fish genes that resist frost, and other plant genes that resist diseases.
2. Increasing the chromosome set count to make the plant larger (which is common even in nature; humans have 2 sets of chromosomes, but plants can have pretty much any number they please, depending on species. Some counts run up as high as 12).
3. Adding or removing growth-medium dependency genes, to ensure the plant is only capable of growing in particular environments. It is possible to make the plant depend on a particular form of fertilizer or additive, and thus prevent it's spread.
All of these techniques have been in existence for several decades and the technology of performing these manipulations is quite well proven.
Plant and crop genetics is a useful field, but they really aren't doing anything revolutionary. They're worth using just for their disease resistance as it increases the consistency of the harvest, preventing the really bad famines where crops could be wiped out by a disease outbreak, but the idea of improving yield beyond the natural limits hasn't had much success.
Y-e-e-e-eah. I mean, using e-coli around food couldn't possibly go wrong. And that whole 'death-gene' thing? Nothing to worry about in the slightest! Not even with rampant outcrossing!
Uh, what? First off, they're taking genes from a non-pathogenic strain of E. coli. Second, they use a genetic screen of both positive and negative growth tests combined with PCR sequencing to ensure that the only genes transferred are the targets. Third, I'm not entirely sure what you mean by outcrossing as it's a term I've never stumbled on in 6 years of genetics education, but if it refers to cross-pollination between GMO and non-GMO variants, GMO crops are typically crippled from doing that. It's an incredibly regulated and safe science.
And let's not forget that a University of Washington study found that “preschool-aged children who had been fed a diet of conventionally produced food had six times more pesticides in their urine than children who had been fed an organic diet.”
Yeah... organic != non-GMO. GMO crops can be "organic" too. All that term refers to is food grown without the use of pesticides, so the study results are rather of the "No sh**?" variety. Furthermore, "organic" food supplies are far more dangerous on the whole because they have a much higher attrition rate (without pesticides, they are subject to all the natural diseases of the 19th century that ravaged crops and decimated yields annually) AND they're capable of carrying bacterial and fungal spores along with other potentially pathogenic pests.
Seriously
Actually, he's quite serious about artificial selection mirroring evolution. Disease resistance is a rapidly-evolving mechanism that can appear in new generations with regularity. As the vast majority of genetic modification to crops is centered around disease and pesticide resistance, it's really not that much of a leap. Especially because it's not like we're "making" genes - any modification done has to come from another type of organism. All these genes exist in nature in the first place.