hlanelee - I would seriously, seriously not post stuff like that on this board. I know ANFO is public knowledge especially after Oklahoma city, but the mods and admins really do not like discussion of anything remotely illegal. They get antsy if people even talk about peer-to-peer file sharing.
Back to the OP - short answer, your friend is right. Long answer…
“Power” isn’t that useful a term when applied to explosives. Practical considerations will almost always trump straight power. Something half as powerful but ten times as cheap may be preferable, for example.
One measure of an explosive’s power is the energy released when it reacts. But energy doesn’t necessarily convert to an energetic blast - it depends on how much of that energy is used to do work. Thermite for example releases a lot of energy on reaction, but the reaction products are non-gaseous - liquid iron and solid aluminium oxide. Thermite is not therefore an effective explosive - too much of the energy is lost as hot non-gaseous product.
Another measure of an explosive’s power is called “brisance” which is a measure of the shattering power. Brisance depends upon the peak pressure developed within the explosive’s body as it detonates, and is determined empirically. It varies with the energy released, the explosive’s density, the detonation velocity etc. Brisance is important if you want to turn a bomb casing into a lot of shrapnel, or break a large volume of rock into smaller bits, but may be completely unimportant or even undesirable if you want to blow tree stumps out of the ground.
Most military explosives are based upon RDX or PETN, which are energetic, brisant, and not terribly sensitive. They have similar performance - RDX is a touch more powerful and less sensitive. C4 is finely devided RDX in a plastic matrix, and so is less powerful because the plastic matrix dilutes the RDX. That is not the point however - C4 is made that way because it is squidgy, which is convenient for a number of applications.
Really big conventional explosions have always used ANFO - a crude fuel-oxidiser mixture. ANFO is a relatively mediocre explosive, but it has a few advantages - it is dirt cheap so you can afford to pile up thousands of tonnes of it, and it is so insensitive that it is not classified as an explosive at all for purposes of transport or storage - it has a lesser designation as a “blasting agent”.
CW’s link to octanitrocubane is fascinating because for a long while it was not thought possible to synthesize this, although it clearly would be a very energetic explosive. Cubane is a highly stressed carbon structure (a cubic lattice with a carbon at each corner and a dangling bond) so the energy of that bond stress is added to the redox energy of the explosive reaction. The high density is also significant as it means the brisance would be tremendous. It remains to be seen if it can be synthesized in any usable quantity or whether it will remain a lab curiosity. Even it is synthesized in usable quantities, chances are you’d do much better with ANFO for any general task, for the money.
Are we at the theoretical limit with chemistry? That’s a tough question. No explosive is as energetic pound for pound as the simple rocket fuel mixtures. A stochiometric hydrogen and oxygen mixture is immensely energetic, for example, but the density is pitiful when they are gaseous. Hydrogen and fluorine are even more energetic, so there may be a way to go yet. On the other hand, explosives have to be practical as well as powerful, and those requirements are to some extent contradictory.
If you want to talk exotic, there are odd states of matter such as metallic hydrogen which may be tremendously poweful explosives. Metallic hydrogen is largely theoretical high density solid state that hydrogen may be able to achieve under enormous pressure. If it indeed exists, and if it is stable once formed, then we’re talking a new family of chemical explosives considerably more powerful than any we have now.