Watson can't cure cancer ... or all the stuff that breaks IT projects A University of Texas audit report last week tipped a bucket on the conduct of a high-profile “Watson to cure cancer” project. The University has criticised the MD Anderson Cancer Center's “Oncology Expert Advisor (OEA) Project”, which since 2014 has poured tens of millions into seeing if IBM's Watson was smart enough to beat …
This was akin to buying fake snake oil, apart from the lack of governance.
IBM was Hollerith originally and made a machine to sort census data in the 19th C.
Watson is simply the 21st C equivalent for "big data". Someone has to put all the relevant data in first. It's only AI because in the last 40 years we have re-defined AI first as "Expert Systems" and now as "Machine Learning".
This is lesson to us all; never buy fake snake oil, INSIST on the genuine article! The consumer/aficionados of fine snake oils know that Virgin means the snake oil was produced by the use of mechanical means only, with no chemical treatment. The term virgin snake oil with reference to production method includes all grades of virgin snake oil, including Extra Virgin, Virgin, Ordinary Virgin and Lampante Virgin snake oil products, depending on quality and amount of venom left in the product. It is known.
ATH+++
Sorry, but an auditor is not there to check on expertise, he's there to check that everything has a trail and the trails all match up to the expenses.
Her complaining about her expertise being dismissed just means that she's pissed that the auditor did not melt before her batting eyelashes and remained steadfast in demanding the paper trail. One would think that, with her level of "expertise", she should have known better.
What this really means is that this project was a vanity project for her, a buddy project for the contractor, and proper controls were deemed "unnecessary" because obviously everything would go fine. Until, of course, it didn't because the controls to prevent that had not been used.
This five year project is the poster child for big projects by big organization with leaders chosen for skills other than getting a great idea to a shipping product. IBM took $40 million, Price Waterhouse $30 million and produced a laboratory mock-up at best.
The leader was connected, with impeccable credentials for healing, for teaching, but NOT for managing a firm with the stated aims of producing a product.
Its not a failure to 'Commercialize' but to monetize. (They are actually two different things)
But MD Anderson does do real research and is one of the top research hospitals for Cancer Treatment.
The problem is that everyone and their brother is trying to cash in on using 'big data' to help fight cancer and in most cases the hype is so far over the top, you will have a hard time talking people down to reality. Everyone is giving money to Cancer research and here's a way to cash in.
The truth is that cancer research is an iterative process and where 'big data' and AI can help is to prune the tree and reduce both the time it takes per iteration and the number of iterations until they find something.
IBM did this in terms of marketing. They used this as a way to suggest how important Watson was. (Its more than just 'big data' but also some analytics packages. ) They used this to keep morphing and selling IBM. The I in IBM doesn't stand for 'innovative' while the M stands for 'marketing'. I be marketing... ;-)
The problem is cancer is actually a group of unrelated diseases that have similar symptoms. They are unrelated as the causes for each is different so the prevent and cure for each is different. How skin cancer, lung cancer, etc. are caused varies so other than surgical removal the treatment will vary. Grasp this fact first.
Just strap Watson, or whatever junky Power Series rack in the box they got lying around and can sell as capital, to the inputs of the wizz-bang IBM Quantum Cloud, harness the energy of thousands of fleeing IT workers to power it, then add in a voice AI input to that, and viola!
"Watson"
Yeah
"Cure cancer! Go! Do it, Boy!"
Okay
"Nice! Now we wait for the punched tape output like on the old Batman TV series, and viola, cancer is cured, me old sons!"
ATH+++
This is close to, if not outright, criminal acts.
”Work performed under an amended scope of work extended beyond the OEA project and intent of funding as approved by the Board of Regents of the UT System (Board)”; and
”Invoices were paid in full regardless of whether contracted services were delivered as agreed upon, and invoice review and approval to process payment to the vendor was not consistently documented as evidence that invoiced services and deliverables were both received and acceptable.”
That last paragraph literally screams "kickbacks."
.... in a post related to IBM's future where someone was saying that Watson had any future beyond jeopardy contests.
"And where are the products? Watson? The perfect death trap for the risk averse IBM: no customer wants to sign off a contract where they will be footing the bill to set up an incredibly complex system with no ROI guarantees whatsoever."
Seems that IBM finally found a customer willing to fall into that death trap
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
