IBM Pocket Watson a Siri-Killer | Hacker News Reader

IBM Pocket Watson a Siri-Killer

bloomberg.com

42 pointsjhull31 comments

I have a feeling this will be as gamechanging as Wolfram Alpha was.

Great technology, but there are already products out there that are okay enough to placate a lot of the incentive for consumers to adapt a non-baked-in application.

I wouldn't use the past tense ("was") talking about Wolfram Alpha. I have spent some time with the free tier of Wolfram Alpha APIs (wrapping the Java code in Clojure) and it seems really useful. I would hope that it just keeps getting better.

I'm not sure what you mean by "gamechanging"

What games has Wolfram really changed out there in the real world?

Wolfram seems kind of nerdy cool, but for me it's mostly been like bar-trivia crossed with nice mathematical function plotting.

That's the joke.

This may or may not be helpful:

http://wolframalpha.com/input/?i=tongue+in+cheek

> A farmer could stand in a field and ask his phone, “When should I plant my corn?” He would get a reply in seconds, based on location data, historical trends and scientific studies.

Historical trends, scientific studies... pfft.

"It's 6AM in the morning, you are lying on the ground and you are two states away from your farm. Also you are a dairy farmer, so the answer is that you should never plant your corn, John."

This is already true and has been for years. Cereal farmers were some of the earliest adopters of GPS and autonomous vehicles. I remember being at a rural show in the early 90's and CASE had a combine harvester for sale that recorded the grain yield per 10 square meters of the field via GPS, and if you bought radio beacons for the corners of the field, it was capable of resolution of about a square meter and semi-autonomous operation (though I wouldn't have wanted to stand in front of it). By this point the software if a major part of what differentiates the various models [0]. These days farming, especially crops is incredible technologically focussed. If you are not tracking crop yield and soil nutrients, you have no way of knowing how much fertiliser, pesticide or seed to spread at each point, which equates to massive waste. Given the margins for farming are incredible slim, and you are often over 5 million in debt, just to finance that years crop, the relative cost of investing in the latest technology is small. By comparison to the integrated systems that tie in everything from soil quality, to transport, to current grain spot prices, this is a toy.

It's not that Watson is not impressive, it's more that I get irritated when journalists talk about farming as if it is merely getting up each morning and riding around on a tractor idea before heading home to the wireless as they haven't yet got anything as new fangled as television. My apologies.

[0] http://www.deere.com/wps/dcom/en_INT/products/equipment/comb...

Misleading title... the article explicitly states they are trying to figure out how to do make it work on mobile platforms and that it could enhance Siri...

Is Siri useful enough to "kill"? From what I can tell it's mostly used for entertainment as people ask questions and laugh at the nonsensical results.

I use Siri every day for dictation, sending emails, replying to texts, and setting reminders. I guess I asked it nonsensical questions on the day I got my 4S... but not since then. I would say: yes, it is useful enough to "kill."

TL;DR: So this thing that has absolutely nothing to do with this other thing is going to kill the other thing because it's going to be so much awesome when it's a thing because as soon as its a thing then you can use it as a thing that's totally unlike this other thing.

Because IBM and smartphone cloud.

One trend I dislike, is the forced use of voice recognition in these cases. A machine as powerful as Watson should not be limited to my mobile device, it should be available via all my computing platforms and should be able to interpret text form.

Siri, for example, can get me David Beckhams age very fast ... It would be nice on my laptop to get access to that type of speed, hit cmd+space "how old is david beckham" and get a result. I understand google can do alot of this, but why limit this technology to voice input only and mobile devices only

This is actually exactly Google's strategy. The Siri-competitive voice search in Jelly Bean is really just an interface to Google search, which has been extended to answer a lot of the same basic questions that Siri can handle (for example, it manages "how old is david beckham" just fine). This means you should be able to get the exact same results by just typing the query into your laptop's Google search box.

You often hear those words, "X killer" but funnily enough Watson is actually a game-changing thing. Even if IBM fails in the consumer market the very fact they're pulling big financial and medical contracts is still just as good for IBM.

It will be interesting to see where this goes and what happens.

It's also not a Siri killer because they are going to sell it to businesses while Siri is mainly a consumer toy.

I love this: two very well capitalized companies competing to create useful AI-enabled systems.

You can reasonably argue that Watson is not a "real AI" but anything that pushes forward AI research is a good thing as far as I am concerned. BTW, I don't spend time worrying about 'evil AIs' running amok like some of the singularity crowd does.

It's so typically IBM to concentrate on Businesses and forgetting consumers with a technology that could even eat into Google's market. They should partner with Microsoft / Nokia and get license fees for every Windows phone. Now that would be some exciting stuff getting people to buy WP8.

If IBM does this and makes a search engine website, they could mop the floor with Google.

Man, I need to know when to plant corn.

Spend enough time on a farm and it doesn't become that difficult to figure out: you guess, based on what you know about the seasonal weather. Watson is never going to know more about the weather than you can get on weather.com.

Actually farmers already pay lots of money for localised weather forecasts exactly because there is a significant amount of information that weather.com doesn't give you.

Not sure what 10 racks of PCs will be able to add to the science of weather forecasting though, given the scale of investment necessary to be a player in that field: http://www.metoffice.gov.uk/news/in-depth/supercomputers (although the Met Office do use IBM...)

Right, I was of course being glib about weather.com. It may then be a useful service for something with Watson's capabilities to narrow down to your specific needs and historical data. I am fascinated by weather prediction, it's an interesting field, and I find I can't trust any detailed weather reading more than 3 days in advance (aside from large weather patterns, which seems readable for about 10 days ahead) so if there is something out there that works longer-range it would be quite amazing to know.

I really do wonder what Watson could add. We always plant corn as close to May 1st as possible as that is what the generations that came long before me determined was best for the conditions here. Neighbours have tried deviating from that date, but the results have always been poor.

Now if Watson could accurately tell me when to sell my corn to maximize profit, that would be pretty awesome.

Where can one read about the machine learning technologies that ibm is working on?

Behold, PoWa!

It can't "kill" Siri because IBM will never be able to integrate it as tightly with iOS due to App Store restrictions. Sorry IBM! That's what you get for building a product instead of a platform.

Android.

They make a deal with Apple and get paid handsomely because of that. An ad-supported app has to be on many devices, and to be run a very long time to be worth $100 million dollars or so (that's probably how much the deal would cost for Apple).

Exactly, this is the IBM of the 21st century we are talking about. They don't make their money by doing business with people, they are all about research and contracting for companies now.

IBM isn't interested in $100M. The article plainly states they're going after billions.

> Finding additional uses for Watson is part of IBM’s plan to tap new markets and boost revenue from business analytics to $16 billion by 2015.

> IBM expects to generate billions in sales by putting Watson to work in finance, health care, telecommunications and other areas.

You must not have read the article. It was not about having a service running two home-screen button taps away from all iPhone users that is also better than Siri. It was about being able to sell a service to businesses with similar AI features as Siri but that provides useful information in specialized fields. This was very obvious if you read any of the article.

> IBM expects to generate billions in sales by putting Watson to work in finance, health care, telecommunications and other areas

> IBM’s path to the mobile-assistant market contrasts with Apple’s. For one, IBM is focused on corporate customers, while Apple is targeting anyone who buys its phones.

> The program will be able to understand oncology well enough to advise doctors on diagnosis and prescriptions, said Martin Kohn, IBM’s chief medical scientist. One iPad application for Watson -- a health-care program developed with a Columbia University professor -- is being used to demonstrate its medical capabilities for prospective IBM customers

Reply on news.ycombinator.com