He had promised to do the impossible -- cryptographically prove he had access to early coins which he did not in fact have. He had carefully developed two scam-proofs of this.

One scam proof was deployed on his blog where he claimed to sign a Sartre text with a key from bitcoin block 9. He carefully did not give the exact input text, only a supposed hash of the text. The hope was anyone trying to replicate would assume their problem was an incorrect source text. And there was proof that the given hash value was signed by the key. But that did not survive internet scrutiny; it was noticed that the signature came from an old bitcoin transaction. Instead of signing the hashed Sartre text as claimed, it signed the old transaction.

The other scam proof was presented in private to reporters and bitcoin developers Jon Matonis and Gavin Andresen. Here Wright supposedly demonstrated the ability to sign arbitrary messages using keys from bitcoin blocks 1 and 9. This scam succeeded: both Andresen and Matonis were convinced. But they were not allowed to keep copies of the evidence, to prevent the trick from being exposed. One speculation is that Wright was able to substitute a doctored version of the Electrum software used for the verification. But without evidence to examine outside of Wright's control, the exact details of the scam are still hidden.

But now Wright is out of ideas. His public-consumption scam failed, and his private controlled-scenario scam can't be more widely replicated.

So this message is his way of backing out, trying as much as possible to save face and keep open the possibility of claiming the Satoshi Nakamoto identity again later.