Here are two key things to consider--

(1) What is the level of company specific risk in each holding?
There are two parts to that:
(a) Some firms have internal diversification, some don't.
A biotech should have an adjustment factor compared to Berkshire, and Hershey might fall in between the two.
(b) Some firms have specific macro risk exposure, like the business cycle, credit cycle, real interest rates, geopolitical or jurisdiction risk, inflation risk, commodity price risk, or even just fad risk.
For either type of these "company specific risk" factors, even using a bad guess is better than making no guess.
One resilient firm might reasonably have the same allocation as several less resilient ones combined.
Note, I do not recommend using price volatility as a scaling factor!

(2) What criteria are used to pick your holdings? Are they random, slanted based on quantifiable metrics, or seat of the pants?
(by "seat of the pants" I mean selected by yourself based on individual characteristics you believe make a good investment, which might or might not be correct in real life)
A slate of randomly picked firms will converge on the performance of an equally-weighted all-market fund quite quickly. There is probably little point in owning more than (say) 40-50.
For non-random selection methods, you might diverge substantially from the average even with a vast number of positions.

Only with some sort of a handle on those parameters can you start to make generalizations about the amount of diversification that is prudent.
For example, I would never put more than 2.5% of my funds into a firm with average company-specific risks that was picked with a quant method that I thought was sound.
And I won't even approach that these days.
However, my allocation to Berkshire is more than half my portfolio, so it's not because I'm a diversification extremist.

Mr Buffett and Mr Munger have commented to the effect that diversification is just a defence against ignorance.
But, it *is* a defence against ignorance, so it has that going for it.
And almost all of us are far more ignorant about the returns of individual firms that we think we are.

I think most people are best served by buying an equal-weight index fund or two, provided they can get an acceptably cheap entry price and a low management fee.
For those who are passable typists but not skilled investment analysts, these days you can skip the management fees by doing this yourself.
Download the list of index constituents, upload it to the broker site and click "rebalance".
Other than the fees, this has the advantage that it lets you create a blacklist if you like. For example, if you want to skip firms based in certain jurisdictions or industries.

Jim