Word came from on-high today that the Amazon.LITERAL slot type has been granted its reprieve. In terms of functionality, this may not be a huge game changer, but it may be a landmark decision in terms of how Amazon works with its development community. We just want to say a few things about that...
Credit where credit is due
First of all, it's only appropriate that we give credit to Amazon for boldly reversing a decision they had previously come to. A lot of companies would be wary of pivoting in this way - to do a 180 is essentially to admit that the previous course wasn't the right one. This really seems to be a victory for Amazon's internal culture of "assert and defend".
It would have been super easy for them to just leave the decision the way it was, and move on. We at 3PO-Labs were among the first and loudest raising the red flag on this issue (blog post, forum thread), but even we had essentially accepted the change and moved on. From a technical standpoint, Alexa Evangelist Michael Palermo was correct in his assertion that you could emulate a lot of the literal's functionality through a custom slot. In our testing we found that it wasn't perfect, but it was pretty good. The thing is, though, that's not really what the uproar was all about...
It would have been super easy for them to just leave the decision the way it was, and move on. We at 3PO-Labs were among the first and loudest raising the red flag on this issue (blog post, forum thread), but even we had essentially accepted the change and moved on. From a technical standpoint, Alexa Evangelist Michael Palermo was correct in his assertion that you could emulate a lot of the literal's functionality through a custom slot. In our testing we found that it wasn't perfect, but it was pretty good. The thing is, though, that's not really what the uproar was all about...
The uproar, in short
Sure, the removal of the literal was going to make freeform matching a little less effective. Frankly, though, a percentage point or two of accuracy (or blindspots caused by the custom slot) wasn't going to be a dealbreaker for most people. But even so, our post on the topic garnered more attention (in the form of comments on the blog post, on reddit, etc) than any other we've written. People were outraged, and a lot of their outrage fell into a few categories:
- Developers were not consulted before the decision was made. The decision may have made a ton of sense internally, but then again we are the engine that populates their ecosystem, so it was important that we were onboard.
- The decision was never announced, it was just noted silently. We might not have even known the change was coming if it weren't for an eagle-eyed member of the Alexa Slack community noticing a footnote added to one of the certification pages. It may not have been the intent, but the optics were that they were trying to slip that one by us.
- It seemed condescending, in its invalidation of our use cases. This was probably the biggest one - the notion that Amazon had decided all of our cool, crazy ideas were not worth supporting anymore. Sure, Dominos and Uber were the ones showing up in the commercials, but that didn't mean that the kooky concept some developer had been working on in his spare time wasn't also a good fit for the platform. For some of us, the thing that really stung is that all of these polished, flashy skills being promoted were skills that weren't innovating or raising the bar with their implementations. Much of the advancement of the Alexa meta was coming from (and continues to come from) the small developers with big ideas; I assert that the Alexa Champions (and those who, frankly, should be Champions - blog post for another day) are the real drivers of platform.
Digging the hole deeper
To make things worse, Palermo's response - while well-meaning and totally accurate - didn't sit well with a lot of us. It seemed like Amazon did not understand the furor over the announcement, and that they were prescribing a solution to a technical problem, as if we hadn't already come up with that approach (for the record, it was a well-known solution in our corner of the dev community). We almost certainly read too much into that post, as it was probably meant for folks who newer to the platform, but at the time it was frustrating that they were addressing the technical problem while not touching the communication problems.
And yet...
Having talked to a few Amazonians about the decision, it seems that the internal conversation never actually stopped. Even while they had moved on externally, there were those inside who kept fighting the good fight. We can't be certain that any of the blog posts, forum threads, tweets, or other feedback actually had an effect on their decision making process, but it certainly feels good to see them address something that the community had been so vocal about.
And really, there's no way to interpret today's announcement in a bad light. There are a lot of good, passionate folks on the Alexa team who want to build cool things, just like most of us. They're still working out the best way to interact with their developers, and what it means to work with their community (instead of occasionally working against it), but as anyone who has been onboard for the last year can attest: their M.O. is leaps-and-bounds better today than it was in January of 2016.
They're on the right track - so here's hoping it's full steam ahead.
And really, there's no way to interpret today's announcement in a bad light. There are a lot of good, passionate folks on the Alexa team who want to build cool things, just like most of us. They're still working out the best way to interact with their developers, and what it means to work with their community (instead of occasionally working against it), but as anyone who has been onboard for the last year can attest: their M.O. is leaps-and-bounds better today than it was in January of 2016.
They're on the right track - so here's hoping it's full steam ahead.
RSS Feed
