After watching it and taking some notes, the biggest problem with this talk is that he loads a bunch of untrue and unfair garbage up front to pander to the local culture of dismissive shitbirding on everyone else outside of the room.
Skip to 22m in to skip all the predictable Chris-Allen-makes-fun-of-everyone preamble and get to the meat of his talk, the substance of which I agree with.
For those who can't stomach it even at 3x speed the way I just powered through, I've got a summary:
- Businesses make bad decisions about the longevity or sustainability of their business practice. They're compose of directors and VPs trying to work on a quarterly OKR schedules for wins to justify their outrageous 200k-250k salary options, so it's a vicious and competitive environment.
- Businesses then try and skip the part about building a sustainable tech org by appealing to, "We hire the best." This is expecting great engineers to parachute in from Valhalla to go to war for you. It's unrealistic.
- Between this style of technical evaluation and absurdly fast product cadences (all free of consequence, no liability expressed or implied), organizations don't feel they need to care too hard about building deep skill in anything but their most core business interest because most of the work is "trivial" (and to his credit, Chris includes a lot of his own work in this).
- Businesses could actually improve infrastructure to eliminate this trivial work, but this requires more sophisticated tools, techniques and designs and these are considered risky.
- This is then used as a justification to remove all technical mentorship and tutelage from the system and only hire specialists.
- And then the industry forms a tight orbit around tools and practices that are deleterious to anything but a short-turnover org like the modern Google.
And for the record, Chris is one of those folks in the community who's made a reputation for being dismissive and cantankerous. It's his natural personality to make even good and insightful points like this hard to watch.
It's too bad the lambdaconf folks didn't give him feedback to just cut his talk in half. If he had started at 22m it would have been a HELL of a talk. Instead he needs to erect an effigy and shove it in a wicker man with bees to appease his (and perhaps his audience's, given the laughter?) love of punching down on folks with less formal CS knowledge.