The one thing the article didn't cover is the patent violation implications of the GM plants growing (and probably being breed for seed-for-sale crops). I'd bet that these particular flowers have patents underlying GM modifications. I found a newer flower color patent with a single google search.  Somebody likely may be paying through the nose for this whoopsie under Monsanto vs. Bowman! 
Are you sure :D
It only takes one idiot putting psilocybin into a dandelion and planting them in the front lawn...
I recant my: did nobody think of Agrajag and the petunias ?
Which is not to say that there aren't any risks at all. But plants and insects are already locked in a millions of years long battle/cooperation. And look how quickly insects have developed resistances to our modern pesticides. Adaption to something new appears to take years to decades, not centuries or longer.
Some insects. And you're very correct that the insect-insecticide battle has been going on for a while - that doesn't mean giving one side a huge leg-up won't have consequences. Genetic engineering can accomplish millennia of evolution in a single generation - the two are not the same.
A few tobacco plants in isolation are going to be protected from insects easily developing a resistance to their natural defenses. We've been growing tobacco plants in massively concentrated fields for 400 years in this country. Nicotine is still a useful insecticide, just as Penicillin is still useful antibiotic.
But genetic reshuffling, recombination, and horizontal gene transfer has been going on en masse in the microbiota we are surround by for billions of years. I think once we start considering introducing truly novel genes, we will have a case for greater caution and reserve. But so far, all projects are of the 'transfer this already existent gene to this other organism', and I think that is a less risky proposition overall.
E.g. how can you know that transferring a gene from maize to petunias is still "sufficiently close" to the level of variation seen in ancient natural processes and evolution?
(Genuine question but also genuine worry about this.)
> Even DNA computers, custom-built for a specific task and then tramped carelessly into wild genotypes like muddy footprints on a pristine floor. Nowadays it seemed like half the technical data on the planet were being stored genetically. Try sequencing a lung fluke and it was even money whether the base pairs you read would code for protein or the technical specs on the Denver sewer system.