Booting up
Sitting on a hard chair for hours isn't the most comfortable thing to do, especially if all there is to do is carefully watch the newly fabricated robot battery charge, and hoping it doesn't explode because of a potential overvoltage. So there I sat, after days of conceiving and tinkering with the shell for which Markus had been preparing the software for ages. I had my doubts about the hardware he had me install in it, particularly in terms of memory, but in the end it was his call.
As I drifted into daydreaming for a while, I figured it would be a great idea to dwell on the pile of plans sitting on the opposite table, silent witnesses of the countless hours spent on this gigantic project: a fully functional android, built complete with more complexity than anyone had ever been able to build. Almost human, as Conor liked to say it, but not quite as human as we are. That comment eventually came out often enough that it had us strive to achieving something even more convincing than what was originally planned.
It actually lengthened the journey quite a bit. The AI was much more advanced than it should have ever been, and knowing Markus he'd gone completely haywire on optimization. Rightfully so: even with the best processing hardware that existed back when we started, there hardly was any margin in terms of processing power to make her work well and treat the deep-learning data accurately enough that it would reach even the first of our initial goals.
So then, as with every technical project, especially those extending this far and wide, the problem wound up data access, mostly the storage and availability thereof. More than once, the specialists in the domain made seminars saying that "common storage mediums are sufficient in terms of performance". The farther we dug into it, the least we were convinced they were right, or that they even knew what the hell they were talking about. But that's a story for another time.
A loud beeping from the battery charger signaling the end of the cycle popped the imaginary bubble I got myself caught into - thankfully so, this new battery technology wasn't exactly the most stable while charging. Its main advantage remained the energy density, and if used properly, the risk was minimal. With it fully charged, the assembly of servomotors and metallic bones was ready to be started for the first time. Without Betty the AI at its core, it would remain an empty shell, but at least I could run the last few tests before Markus arrived with the final version for the software.
The mobility and stress tests passed, it was time for the pièce de résistance to show up. Given the specific nature of the hardware, a few hours were required to fine-tune the last adjustments. Then finally, we heard her voice: "Booting up."
As I drifted into daydreaming for a while, I figured it would be a great idea to dwell on the pile of plans sitting on the opposite table, silent witnesses of the countless hours spent on this gigantic project: a fully functional android, built complete with more complexity than anyone had ever been able to build. Almost human, as Conor liked to say it, but not quite as human as we are. That comment eventually came out often enough that it had us strive to achieving something even more convincing than what was originally planned.
It actually lengthened the journey quite a bit. The AI was much more advanced than it should have ever been, and knowing Markus he'd gone completely haywire on optimization. Rightfully so: even with the best processing hardware that existed back when we started, there hardly was any margin in terms of processing power to make her work well and treat the deep-learning data accurately enough that it would reach even the first of our initial goals.
So then, as with every technical project, especially those extending this far and wide, the problem wound up data access, mostly the storage and availability thereof. More than once, the specialists in the domain made seminars saying that "common storage mediums are sufficient in terms of performance". The farther we dug into it, the least we were convinced they were right, or that they even knew what the hell they were talking about. But that's a story for another time.
A loud beeping from the battery charger signaling the end of the cycle popped the imaginary bubble I got myself caught into - thankfully so, this new battery technology wasn't exactly the most stable while charging. Its main advantage remained the energy density, and if used properly, the risk was minimal. With it fully charged, the assembly of servomotors and metallic bones was ready to be started for the first time. Without Betty the AI at its core, it would remain an empty shell, but at least I could run the last few tests before Markus arrived with the final version for the software.
The mobility and stress tests passed, it was time for the pièce de résistance to show up. Given the specific nature of the hardware, a few hours were required to fine-tune the last adjustments. Then finally, we heard her voice: "Booting up."
Comments
Post a Comment