I can't find anything that looks like it googling the Ganguli translation, online here. That doesn't necessarily mean it isn't there. Could be I just didn't think of a suitable search string to match this translation.

From the Ganguli translation of The Mahabharata Book 14: Aswamedha Parva, Book L, at http://www.sacred-texts.com/hin/m14/m14050.htm:

"Well, I shall now tell you something more. With even a thousand (explanations), one that has a bad understanding succeeds not in acquiring knowledge. One, however, that is endued with intelligence succeeds in attaining happiness, through only a fourth share (of explanations)."

Aha! So it's Ancient Wisdom with modern rephrasing for greater impact!

Good lord, it actually is from the Mahabharata, if somewhat rephrased! Excellent find, ed.

But it's only one thousand, not ten thousand. Brahma never extended his studies to giving a foolish man TEN thousand explanations. That work remains to be done. Perhaps he would understand after 3 or 4 thousand.

The Buddhists learned about exponentiation and then went nuts.

"-- If for every grain of sand in the Ganges river there were another Ganges river, and if for every grain of sand in all those Ganges rivers there was a whole universe full of gold and jewels, would that be a quite a bit?

--Quite a bit indeed, O Honored One."

From this transcription:

0140490171/.sahasreNa.api.durmedhaa.na.vRddhim.adhigacchati./
0140490173/.caturthena.apy.atha.amzena.vRddhimaan.sukham.edhate.//

At least I hope my mediocre vocabulary didn't betray me and that's it. Is there a Unicode canon someplace? The Devanagari ones I found used funky fonts.

Thank you!

Found it. (Thanks to the Wikipedia entry.)

14049017a सहस्रेणापि दुर्मेधा न वृद्धिमधिगच्छति
14049017c चतुर्थेनाप्यथांशेन बुद्धिमान्सुखमेधते

MMcM, in the second shloka, I'm reading:

caturdhenaapyamzena buddhimaansukhamedhate

and not vrddhima... and so on. True, bu and vr are fairly similar. Whose error is this? Which reading is correct?

This page describes the process of developing the Cambridge text. It started with the Kyoto (quoted above) and corrected it. In addition to some transcription errors, Tokunaga had undone the sandhi to faciliate searching. Smith didn't think that a good idea and it was redone.

It looks like the text is stored in some canonical form and delivered in various encodings (via CGI scripts maybe). Here is the same text in their ASCII format (again a slight different system -- capitals instead of doubled for long vowels).

14049017a sahasreNApi durmedhA na vRddhim adhigacchati
14049017c caturthenApy athAMzena buddhimAn sukham edhate

Since differences are conscious repairs (and not alternate sources), I think the answer to your question is "the second one".

OK, here it is in (Unicode) IAST romanization:

17
sahasreṇāpi durmedhā na vṛddhim adhigacchati caturthenāpy athāṃśena buddhimān sukham edhate

The spelling is as MMcM's above, but then I found it by converting that and searching for it. It's from the same (Cambridge) text, anyway. Front page here.

So for the benefit of those who don't speak sanskrit -- is it 1000 or 10,000? And are they
easily mistaken for each other?

It's a thousand; I assume the larger figure was introduced for effect in the English version. Certainly not a misunderstanding of the Sanskrit.

Better Lisboan than Lisboner, I guess.

Lisboan. Well I never.

Of course, our most famous c#sino here in Macau is the "Lisboa", known in Chinese as the 葡京 (Portuguese capital). (Why is your software censoring the word cas#no?)