Chips do close to nothing for non turbo applications. All it can do is take the safety margin for air flow and make it slightly more lean, it cannot push more air into the engine. This gives you a slightly more efficient burn, as long as everything is working perfectly well. If something goes wrong, and it will "just because" you will put a bit more heat/ping/detonation into your engine than it was designed for. That will lower the longevity of the engines.
Another issue is that every engine is different. When they program the ECU it is pretty lax and can cope with engine differences. These chips cannot.
So can they work? Sure, probably 10hp increase if you are lucky. If you get a customizable one and use a dyno it is probably actually safe, if the guys doing the programming know what they are doing. There you might see a bigger midrange bump as they match everything up.
If you port and flow match the heads, put on headers, change exhaust and that sort of thing, then yes you could see huge differences. The stock ecu probably cannot deal with those sort of changes.
In a turbo they just keep the engine running a bit rich and tell the system to not dump air until it gets to higher pressures, increasing the available boost. Works great and you can see significant increases in hp. Since most sane places have turbo diesel engines, thats why you hear of chips so often. Here in the USA... nope, all normally aspirated gas engines.
BTW: if a deal sounds too good to be true... Chips normally run in the 800-1000 range due to the cost to break the encryption/systems used in the stock ECU, not to mention testing and programming. For more mass market applications where they might sell 10-20k chips, the cost will be less.