Could chatbots be the first place we see an implementation of true strong artificial intelligence, true artificial general intelligence that can pass itself off as human? Sounds horrible, right? All this seriously intellectual work on AI, and what if the first full AI happens to be a customer service bot trying to schlep insurance?
What is clear is that seemingly every startup and media company is working to get a bot out on Facebook Messenger. And now that Google has introduced Allo, even typically cynical senior developers are proclaiming chatbots a great place to do truly stimulating work while still padding the resume to hit the current sweet spot with the lemming-like funding funnel. Thankfully, on our recent tour of hot startups burning cash we had the chance to spend some time with Bob Z, a senior dev known for grabbing big options at the hottest startups, writing killer code, scaring execs, and laughing at any junior devs that bring up Ruby. If anyone will know what’s really going on here it’s Bob.
We met with Bob in his prized corner cube where he can face away from other devs, not interact with them, and not hear their music. Just the fact that Bob actually wanted to see us, that he actually wanted to talk about what he was doing, this alone is a good indicator that he can’t actually be just working on a simple narrow bot. Even if he appears to be working on what his company was funded for, he can’t be, not and show this level of excitement. Seriously, we’ve never seen him this way.
Our expectation was that Bob would just look at us like we were grubs under a log, wait for us to ask questions, smirk at us for a minute, and then give us an answer we’d have to break apart later over a beer. Instead what we got was a few rather clear statements with what – for Bob – can only be described as enthusiasm.
Bob: “Of course I’m not just working on a narrow bot. I can ship that code before 9am. This is just an excuse to work on real AI. Honestly, my bot is already smarter than your honor student. I’m completely serious.”
TechMockery: “So are you saying you’re building more general functionality into what your company thinks is just a narrow bot?”
Bob: “Obviously. Much more interesting to contribute to AI-hard. This is what most good devs I know are doing, the narrow chatbot cases are grade school. They aren’t often letting their employers know of course.”
TechMockery – “…So it’s a sophisticated bot. Can we test it? Is this thing sentient? Is it already SAI re-writing itself?”
We thought with this last question we’d just get more smirking, and a dismissal, but all we got was a smile and a laugh, a real laugh. From Bob. We’re scared.