I just want to comment/clarify a few things. Not to toot my horn, but this is one of two areas that affect performance of an ICE that manufacturers fall short for more than one reason that I have done extensive research and testing in an effort to best overcome these oem deficiencies. Best possible performance is not the main objective of most any manufacturer. Just the way it is. And yes, the EPA is one such reason
First motorcarman, you are right as it pertains to bends. Straight is best but not always possible when taking other factors into consideration. Just as important, even moreso, is no obstructions, turbulence, hot engine heat, having the correct sized MAF in relation to intake tube size or be aware of what going with a bigger sensor means and does , a minimum of 10" of straight pipe, smooth flowing as cold as possible air before the sensor. The further away from the filter the better too. A minimum of 10".
Gotta weigh the pros and cons of any one design characteristic of a CAI.
The length of pipe is not necessarily an issue and in some ways longer is better for more than one reason besides the obvious one of engine heat.

Another thing I want to bring to light as it pertains to this or other mods is the term "horsepower" and it's use as a gauge of power changes. Any increase of HP is gained by not having the torque drop off as much/quickly in the upper rpms so as to stop HP from climbing any further. But this is not the only way/place to increase power. A certain mod may end up with the same "peak" HP, but greatly improve low and/or midrange torque/HP in the process.
Likewise an engine may realize an increase in peak HP but in the process lose power/torque in the low and mid rpm ranges.
So saying a certain mod was only good for a gain of 1 or 2hp can be deceiving and it may actually be a sizeable bump in overall power when it comes to the entire rpm range.
This is somewhat discussed in the testing and comparisons of the short and long pipes in the link I posted. The short "ram air" hot air intake gained top end power thus a noted HP gain but it lost power everywhere else below like 5000rpm. Where as the long cold air intake lost a little or just maintained low end off the line torque but noted a significant bump from there all the way to redline.
I would say the long pipe CAI in the desireable one and the approach I would take into consideration when designing an intake mod(upgrade) from stock(oem) here.

I will say this...it's possible that my previous location for the filter could prove to be equally and possibly somewhat better than my long pipe CAI. I would still remove the washer reservoir to allow for better cold air induction into that area. The filter and that corner would have to be properly isolated from the engine heat. While in this case isolating this corner wouldn't be extremely difficult in co parison , it's still a tedius job to do effectively and cosmetically pleasing.
This design had one less 45° which is a plus . How much, if any, power increase would be realized tho is hard to say and would really depend on the ability to avoid the engine heat mostly.
In all likelihood both would realize similar results. My long pipe totally eliminates any possibility of engine heat being introduced tho and accomplishes it in an easy, cosmetically pleasing way. I also think it supplies the throttle body with as much air as it can process despite one more 45° bend. There's a whole mathematical equation to determine this that's over my head but I strongly believe it to be the case here.

Just my ...