An Epidemic of AI Misinformation

It's the bullshit process by which people get funding. Make something that has some potentially interesting engineering value, hype it up to hell and back, and funding agencies, investors etc. are more likely to back you. Probably this wasn't everyone's first plan all along, but when people who make a marginally better hot-dog / not-hot-dog classifier put so much spin on it, then everyone else has to just to remain visible. Moreover, people have to publish findings before their competition does, meaning sloppier and less interesting work. It's the snake-eating-its-own-tail plague that impacts so much of academia.

Good article, but it overstates its case.

>In 1966, the MIT AI lab famously assigned Gerald Sussman the problem of solving vision in a summer; as we all know, machine vision still hasn't been solved over five decades later.

It certainly took five extra decades, but it would be a massive shift of goalposts to say the problem of vision hasn't been sufficiently solved today.

>In November 2016, in the pages of Harvard Business Review, Andrew Ng, another well-known figure in deep learning, wrote that “If a typical person can do a mental task with less than one second of thought, we can probably automate it using AI either now or in the near future.” A more realistic appraisal is that whether or not something can be automated depends very much on the nature of the problem, and the data that can be gathered, and the relation between the two. For closed-end problems like board games, in which a massive amount of data can be gathered through simulation, Ng’s claim has proven prophetic; in open-ended problems, like conversational understanding, which cannot be fully simulated, Ng’s claim thus far has proven incorrect. Business leaders and policy-makers would be well-served to understand the difference between those problems that are amenable to current techniques and those that are not; Ng’s words obscured this. (Rebooting AI gives some discussion.)

It takes significantly longer than a second to actually understand spoken conversation (rather than provide a conditioned response or match against expected statements, both of which computers are fully capable of doing).

>I just wish that were the norm rather than the exception. When it’s not, policy-makers and the general public can easily find themselves confused; because the bias tends to be towards overreporting rather than underreporting results, the public starts fearing a kind of AI (replacing many jobs) that does not and will not exist in the foreseeable future.

Robotic manufacturing has already eliminated massive swaths of high paying jobs. Likewise, software has eliminated massive swaths of data entry and customer service jobs (with software being a particularly poor replacement for the latter, but still being put into widespread use to cut costs). And contrary to beliefs that new jobs will be created in IT, software is able to massively eliminate low skilled tech jobs as well, as e.g. automated testing did to India's IT industry.

As with existing jobs that have been automated away, companies won't need generalized AI to eliminate many more jobs. Many jobs don't rely on unconstrained complex deduction and thinking, and will be ripe for replacement with deep learning algorithms. And we can be reliably assured that corporations will engage in such replacements even when the outcomes are not up to par.

That's probably a part of it. Note how the Google TPUs aren't for sale. If you want them you have to use the Google Cloud. The cloud is expensive and slow ... I think everyone is shocked when they first see perf numbers coming off Azure.

I don't know if GCE is better, but the temptation to overload the hardware is always there: hardware rental is fundamentally a business with low barriers to entry. Anyone can buy some machines, bring up a Kubernetes or OpenShift cluster and start renting it out. So the big 3 are always looking for proprietary advantage and dedicated AI chips are something other firms can't easily do at the moment, making it a good source of lockin.

Do many people need it though? Deep learning is pretty useless for most business apps, unless you happen to need an image classifier or something else pre-canned. Classical ML is often sufficient, or better, human written logic. The latter can be explained, debugged, rapidly improved and in the best case requires no training data at all!

Apologies, but I see a lot of “trash” published in arxiv (also some real good stuff) Peer review could fix some things but would also show some other things down

I think this is what happens when you have various stakeholders and some are driven by KPIs such as reach / views. To that end those specific stakeholders (content writers / marketing) will bias towards sensationalism. That's not to excuse the organizational culture which allows for this behavior / outcome. There does need to be course correction.