Bottom line: speech recognition in the general case (more than a few predetermined words) is only as good as the 1) acoustic model (which utterances were heard), and 2) language model (how do we group the utterances into words).
This requires massive amounts of labeled data. This is why Nuance is king and few others come close - the amount of labeled data necessary to catch up is astounding. Not to mention a patent minefield to navigate.
This is unfortunately one field in which open-source alternatives face real obstacles and won't be viable in the near future.
and it didn't seem to have a log of words labeled each by timestamp offset into the audio recording - which is the vital part for training a recognizer. Am I missing something?
It is very interesting but unfortunately just appears to be too much hassle for sane people to tackle (although it'd be extremely worthwhile if someone would innovate in this space and lower the barrier to entry - most are using commercial acoustic models with the FOSS software)
Google already does not fulfill subpoenas for email seizures in many cases due to the decision in Warshak - their contention is that they cannot be sure if the user resides or uses their email under the jurisdiction of the Sixth Circuit.
It's puzzling to think about why Warshak wasn't appealed to the Supreme Court (SC), where a decision would have federal instead of regional consequences. The Justice Department most likely felt such a move could easily end in the SC siding with the Sixth, ending the free lunch on electronic communications seizures that law enforcement currently enjoys.
Google's particular problem, in my humble opinion, is that the protections do not apply to non-content portions of the communication. For example, the government does not need a warrant to require production of the smtp logs, just as they don't need a warrant to see what phone numbers you called whereas listening to those same conversations does require court authority.
Google unlike most other email providers is going through the content of every email (I assume prior to its being read by the user) and indexing its contents for the purposes of determining relevant advertising (and whatever else they do with that info about which I admittedly know very little).
I would wager that those indexes might fall under the "log" rather than "content" aspect and therefore their production would not be subject to warrant if they are keeping it stored somewhere; but, that's for someone receiving a higher pay grade to determine.
Cert for Warshak was not sought by either party most likely because the outcome of the appeal was largely in the govt's favor. Most of the convictions stood while only some were remanded. So I imagine it wasn't in either parties' best interest to roll the dice again. Finally, the precedent established in Warshak is applicable in the other horizontal jurisdictions. I would find it hard to believe any of the other appellate courts would go against its sound reasoning. But I guess the point stands that if you want to be safe you should ensure your servers reside in the jurisdiction of the Sixth!
This is the most informed post here so far. A couple of important points to add.
The problem is 2703(f). It allows a government entity to request a snapshot of a user's account "pending the issuance of a court order or other process", meaning that 180 days later, the government can just issue a subpoena - they get their foot in the door and then come back later with the lowered bar, and only need a subpoena. Same effect as not having warrant protection.
http://www.law.cornell.edu/uscode/text/18/2703
It's important to point out that the 90-day notification can be renewed indefinitely, pursuant to 2703(f)(2), which is probably a breach of what the 4th Amendment intended.
The concurring opinion authored by Judge Keith was directly on point of your second paragraph regarding the indefinite renewal constituting an illegal wiretap. It concludes that 2703(f) is vacuous as it is clearly fails constitutionality for this reason. [0]
I'm sure the irony that such an important case is based on the illegal money laundering of a company peddling a penis-enlarging drug scam will not be lost. Remember Enzyte and the smiling Bob ads? I'm still amazed how easily they got so many merchant accounts given their chargeback rates but I guess that's what fraud will get you.
This requires massive amounts of labeled data. This is why Nuance is king and few others come close - the amount of labeled data necessary to catch up is astounding. Not to mention a patent minefield to navigate.
This is unfortunately one field in which open-source alternatives face real obstacles and won't be viable in the near future.