As noted in earlier threads, this is a false choice. The alternative to "TSA doing everything they do now the way they do it now" is not necessarily "no screening." We recently established that no one here is suggesting that there be no screening, only that the screening be reasonable and in proportion to the threat. Which brings me to:
But what you have "at your disposal" is $7 BILLION a year with requests for more (yet still not enough, obviously for the magic liquid-explosive-finding x-ray). How much is enough? When do you say "we could do something else but the cost is too high?"

There are three things to be considered in any restriction on items and the associated screening: (a) what is the actual risk, (b) how intrusive/slow/expensive is screening for it, and (c) to what extent will people be inconvenienced by not being allowed these items.

Take two extremes: a Wily Coyote ACME-brand bomb is (a) dangerous and (b) easy to find with x-ray or WTMD. As for (c), it's hard to envision anyone who legitimately needs to carry on an ACME bomb and will be inconvenienced if they're banned. This is an obvious case of "should not be allowed."

On the other hand, consider liquids. Every bomb is a threat, every gun is a threat, but most liquids are harmless. Just for now, I'm going to pretend that liquid explosives are a viable threat. But compared to guns or Wily Coyote bombs, liquids have not proven, historically, to be a huge risk. So on part (a), liquids score low (maybe not zero, but low). For part (b) finding liquids is more difficult so the screening now has to include a patdown or nude-o-scope. People have to answer medical questions about their legitimate liquids, risk contamination through testing, risk having medicines or baby food confiscated by your colleagues who get it wrong, decide to leave important medicines at home because they're worried about getting in trouble, etc. In short, screening for liquids is far more intrusive, expensive and slow than screening for guns or bombs. As to (c), many, many people have legitimate needs to travel with liquids, which means that you need to allow exceptions (which negate the restrictions) and also that many people with legitimate intentions (bringing a bottle of wine back from overseas, for example) are inconvenienced.

Instead of TSA increasingly restricting every single thing that could possibly be a "viable" threat, they need to weigh up factors in this way. This is what I meant by saying screening should be proportional to the threat. Is the increase in cost (not only to TSA but to the passengers, in real money and in time), intrusiveness and inconvenience for law-abiding citizens justified by the extent of the threat? It is clear that TSA NEVER does such an analysis, they just respond to every "viable" threat in the same way.

When I leave my house, I turn off the oven, stove and steam iron. But I don't unplug all the electronic devices which are on standby and I may leave a light or two on, even though there is a risk those things may cause a fire. Why? Because the risk of fire from the TV or DVD player or 100W bulb is very very very low compared to a fire from the gas cooktop. And unplugging everything in the house is inconvenient. I make an assessment of risk versus the difficulty of eliminating that risk. Similarly, I lock the front and back door. If I took the TSA approach you suggest: "we should try to prevent what we can with all means at our disposal" I would have to hire armed guards to patrol my house day and night. Some cities in the world, that's actually a good strategy, as the risk outweighs the cost. But not where I live.
Probability is one of the things that needs to be taken into account, and the one that TSA ignores. There are literally billions of things/events that could be dangerous 1 time out of 1,000,000,000,000,000,000,000, but the risk is not sufficient to justify screening for them. (For example, there's a very tiny but non-zero risk that all the polyester seat covers on the plane could spontaneously burst into flame.) At the same time, TSA ignores obvious risks (cargo, employees) because it's "too hard."
Apply the rationale above. (a) Guns are very effective if you want to hurt or threaten people. (Compare that to liquids, which, per se, are not weapons.) (b) Guns are easy to screen for with x-ray and WTMD. And air travellers have decades of experience of restrictions on guns. (c) People who need to transport guns should have safe methods to do so without carrying them on board. (Note - TSA needs to improve luggage security so this is true.) Sum: yeah, screen for guns.
Actually, I think you should, at least for box cutters and pocket knives. (Machetes and swords should probably still be forbidden.) (a) risk from small knives is fairly low, but (b) screening is fairly easy but (c) lots of us would like to have our Swiss Army knives back.
The CYA approach is a bad way to do security. All it would take is one TSA employee getting an explosive through security and blowing up a plane, and we'll all be raising hades about how you aren't screening your own people. TSA is willing to accept that risk because it would be inconvenient for you. When it comes to inconveniencing us, however, TSA is not willing to accept ANY risk, no matter how small.
As noted by others, the "oops, oh well, we're only human" only applies to the TSA. Passengers don't get the same consideration.
Only if (a) someone can show that they've been harmed by your processes and (b) has the money and time (and more money) to take it to court. It'll happen, but not right away. And "we'll do this until someone says it's illegal" is not a responsible way to run a gov't.
Quantity of training is not the same as quality of training.
If you get to judge all passengers (more than 99% of whom are law-abiding innocent people) in the light of Richard Reid and the 9/11 perpetrators, we get to judge all TSA screeners on the basis of the far more than 1% who steal, yell, lie, make up rules, mock, threaten and ask D-Y-W-T-F-T. The TSA can't simultaneously present itself as infallible and above scrutiny, then hide behind "nobody's perfect/room for improvement" when found out.