I think he's saying that you can measure how much energy the food contains but not how much energy each individual will successfully absorb and metabolize.

Nah,, that's the funniest attempt at dissing someone that said something you don't understand I've ever seen.

Calorimeters do a specific job. That job is not the same as digestion and metabolism. Not all foods "give up" calories in the same way, and no foods do so in the same way as inside a calorimeter.

Measured calories via calorimeter are indeed accurate with exactly what they measure, i.e. The exact food that is placed into them.

What a calorimeter can't do is guarantee that everything put into it is the same.

The more complex the substance is, the more variation there will be between measurements of different batches of that substance. Something like refined sugar is going to give the same results reliably because there's just not that much variation. Same with refined fats and proteins. Once you get simple enough, the results vary by so little as the be meaningless.

Put two bananas in the same machine, the variance will be greater than that of simpler materials. Is that variance enough to matter on a practical level? Not usually, but it can be.

But, that variance is still there, and the range of possibilities is enough to be significant when calculating what you might slap on a nutritional level of a given food.

Hence, the results aren't accurate in the sense that they can be reproduced in a precise way. There's just too much natural variance in foods, even carefully prepared foods.