He gives an anecdote on how hard it would be to create just one new protein by chance
His anecdote ignores protein functional redundancy. The key thing about a protein molecule is its physical shape which results from the sequence of amino acids which results in turn from the sequence of nucleic acids.
The thing is there are many many possible sequences that result in an identical molecule. Taking just one example of Cytochrome C. The number of possible amino acid sequences that would result in a functional Cytochrome C protein molecule has been calculated to be a billion times larger than all the atoms in the known universe. And each one of those amino acid sequences could be built from vast numbers of alternative sequences of nucleic acids.
All of these billions of billions of possible sequences results in a perfectly functional cytochrome C molecule.
Now consider all the orders of magnitude of other possible sequences that would result in a cytochrome C molecule that works a little bit less efficiently than those we find in extant species.
His argument is like somebody looking around a stadium of 100,000 people and asking what are the odds of these exact people being right here, each in their exact seat at this precise moment. They forget that countless other contingencies would have also resulted in a full stadium.
Prof Gelernter to needs to stick to computers.