That's fairly neat. If I'm understanding it right, it's a mechanical dial-turner which clips onto a stove burner's control, and is combined with a digital thermometer which reports the current temperature in a dish on said burner; the dial-turner targets a temperature in a negative feedback loop.
(In retrospect, I'm surprised we don't have 'smart' stoves which pair up with dishes with thermometers built in to allow seamless temperature specification. Would work very well with induction stoves, I think, since they can adjust the heating on/off much faster than a regular gas/resistance-heating stove.)
I'm one of the founders of Meld and I'm happy to answer any questions you might have about our product or precision cooking and temperature control in general.
We were as surprised as you that this doesn't exist already. A traditional stove is an amazingly capable and flexible device. By adding some simple smarts to the burner control it's possible to do amazing things.
Induction stoves are good, but you might be surprised to learn that gas is even more precise. With gas you basically have infinitely fine control over the flow of energy to the pan since a gas valve is a linear analog control. Honestly, though, whether you have a gas or electric stove we can generally keep temps to within one degree (F) of a given target.
I totally understand your concern. That's why we NEVER turn on the stove for you. All modern stoves require you to push the knob in to move it from the off position. The Meld Knob is mechanically incapable of defeating this safety interlock.
I'm an engineer so when people ask if something is possible my honest answer has to be that anything is possible.
Would you allow me to rephrase your question? I'd rather answer the question, "What is the likelihood that the knob can be installed incorrectly or that the mechanical interlock will fail in a way that would allow the knob to unexpectedly turn the burner on when not intended?"
My answer to this question is that it would be EXTREMELY improbable that this would happen. I say this for the following reasons. 1) The part of the knob that allows the user to push the shaft inward to override the safety interlock is a completely separate piece from the part that turns the knob; 2) Both of these pieces are made of smooth, lubricious plastic such that it should be very difficult to install the knob in a way that they would bind; 3) assuming the user did install the knob in a way that made this possible it would be very obvious to the user that things are not right; 4) the knob software is written in a way that the knob will refuse to move from the off position without physical intervention from the user.
What I've described is generally called a defense in depth approach. You put a bunch of safety mechanisms in place such that all of which have to simultaneously for a bad event to occur.
(In retrospect, I'm surprised we don't have 'smart' stoves which pair up with dishes with thermometers built in to allow seamless temperature specification. Would work very well with induction stoves, I think, since they can adjust the heating on/off much faster than a regular gas/resistance-heating stove.)