Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Jefferies gives IBM Watson a Wall Street reality check

techcrunch.com

Post reply on HN
Print view View original

71–80 of 295 posts

Re: Jefferies gives IBM Watson a Wall Street reality check

#71

Post by PKop »

filereaper wrote: post #62
xienze wrote: post #25

> Jefferies pulls from an audit of a partnership between IBM Watson and MD Anderson as a case study for IBM’s broader problems scaling Watson. MD Anderson cut its ties with IBM after wasting $60 million on a Watson project that was ultimately deemed, “not ready for human investigational or clinical use.” Well, can't say I'm surprised. I used to work on that project a few years ago, basically the idea was that Watson…

So if I understand correctly, there was dirty/malformed data which was difficult to interpret, and when sent to a ML algorithm not tuned by a domain expert led to bad results. This applies to all ML work, why is Watson exempt from it? Conversely if this the canonical failure case, why's everyone so harsh on Watson?

Perhaps because they charged $60 million for it?

Re: Jefferies gives IBM Watson a Wall Street reality check

#72

Post by retreatguru »

matt4077 wrote: post #44
abtinf wrote: post #27

Earlier quoted context omitted.

Disclosure: I work in developer advocacy at IBM. Thank you for sharing your view. If you'd like to talk more about how we can better message our offerings, please feel free to reach out. I'd love to learn more about the kinds of problems you are trying to solve. Edit: yes, I know, this is a bland response to the parent. But change needs feedback. Edit 2: thanks matt4077

Here's some feedback: sentences like "getting feedback is essential toward driving change" are the reason nobody likes IBM. "Change", as used here, is a noun based on the verb "to change". If your sentence calls for a verb, you can just go back to the original "to change". Something isn't essential "toward" something else. It is essential "for" something else. "is essential" is unnecessarily passive. Use active verbs…

I loved your dissection of classic corporate-speak!

Re: Jefferies gives IBM Watson a Wall Street reality check

#75

Post by filereaper »

xienze wrote: post #65
filereaper wrote: post #62

Earlier quoted context omitted.

So if I understand correctly, there was dirty/malformed data which was difficult to interpret, and when sent to a ML algorithm not tuned by a domain expert led to bad results. This applies to all ML work, why is Watson exempt from it? Conversely if this the canonical failure case, why's everyone so harsh on Watson?

> This applies to all ML work, why is Watson exempt from it? I'm not following, Watson was most assuredly not exempt from it. > Conversely if this the canonical failure case, why's everyone so harsh on Watson? Because they were attempting to use Watson to help make potentially life-and-death decisions?

>> This applies to all ML work, why is Watson exempt from it?

>I'm not following, Watson was most assuredly not exempt from it.

Apologies, I meant it as not having clean data with badly tuned models should be obvious root cause of failure to the analysts who are familiar with this industry. As in it may take more work over time to improve the situation, and that this problem occurs commonly in other areas of AI/ML as well.

>> Conversely if this the canonical failure case, why's everyone so harsh on Watson?

>Because they were attempting to use Watson to help make potentially life-and-death decisions?

Touche. I have no arguments here.

Re: Jefferies gives IBM Watson a Wall Street reality check

#76

Post by throwaway91111 »

Swizec wrote: post #70
zoeysaurusrex wrote: post #61

Earlier quoted context omitted.

I'm happy to be harsh because I've had years and years of experience dealing with IBM. Their tools are trash for the cost they want. The majority of their offerings are overengineered and prone to failure in a production setting. They are pros at selling to non-technical management, all the while IT employees get stuck dealing with the aftermath. We ended up getting stuck with the IBM version of work item management,…

"Nobody ever got fired for buying IBM"

Please, let this quote be ironic.

Re: Jefferies gives IBM Watson a Wall Street reality check

#77

Post by ballenf »

save_ferris wrote: post #12

I don't think this isn't limited to IBM, my partner's PE firm recently hired a small consulting group touting a "revolutionary, AI-driven" real-estate analysis product that has zero AI whatsoever. It's basically a custom spreadsheet tool that they're claiming to be building AI on top of as they consume company data, but for a few hundred grand per year, they have a basic CRUD app on Azure with a reporting tool using…

Later this year I predict they'll relaunch the product as being built on the blockchain.

Problem solved.

Re: Jefferies gives IBM Watson a Wall Street reality check

#78

Post by throwaway91111 »

pdkl95 wrote: post #24
bpodgursky wrote: post #11

Earlier quoted context omitted.

To be fair, the "technical" person at a conference does not mean "engineer". It's like a quarter tier past pure-salesperson. Obviously not an impressive showing overall, but I don't think it should be used as a judge of their technical talent.

I understand that the people sent to a conference may not be highly knowledgeable about some of the companies products. However, if the company isn't going to at least train them for easy questions like why anyone should buy the product featured prominently in the booth's advertisements, then what was the point in renting vendor space? [1] At a minimum, give them a handbook with the official answers for common questi…

Yea, watson's distinguishing features can't be "magic beans".

Re: Jefferies gives IBM Watson a Wall Street reality check

#79

Post by MaysonL »

Swizec wrote: post #70
zoeysaurusrex wrote: post #61

Earlier quoted context omitted.

I'm happy to be harsh because I've had years and years of experience dealing with IBM. Their tools are trash for the cost they want. The majority of their offerings are overengineered and prone to failure in a production setting. They are pros at selling to non-technical management, all the while IT employees get stuck dealing with the aftermath. We ended up getting stuck with the IBM version of work item management,…

"Nobody ever got fired for buying IBM"

And IBM often tried to get people fired for not buying IBM – a former colleague was once the CIO of the only Fortune 100 company that didn't have a corporate IBM mainframe (one or two of their subsidiaries did, but the conglomerate mother ship did not, only Univac). IBM had quite a few golfing conversations with the CEO trying to get him fired. [This was back in the '70s].

Re: Jefferies gives IBM Watson a Wall Street reality check

#80

Post by tomjakubowski »

blazespin wrote: post #60
_delirium wrote: post #17

Earlier quoted context omitted.

"Watson" made more sense to me when I realized it was just a branding term for any AI-related product, API, or service offered by IBM, which vary widely in functionality. That fact is not that hidden [1], but their marketing doesn't exactly emphasize it. If you view it that way, "Watson" should just be compared against any other similarly broad AI-as-a-service offering with strengths and weaknesses, not treated as so…

It might be more honest if they said "IBM Data AI Division."

"IBM DAD"? :)

More seriously, do you think Microsoft would have been more honest if they'd called Windows "Microsoft Graphical Multitasking Operating System"? Or if BMW had gone with "BMW Better Performance And A Cooler Badge But It'll Cost You" instead of "BMW M"?

Brand names exist for a reason.

Post reply on HN