Anyone still with IBM able to give a view of how WatsonX is really doing in the market. Is it selling? Any issues? Layoffs?
14 replies (most recent on top)
I'm just an AI novice (aren't we all?), but it seems to me that after reading the CP4D and watsonx briefs that IBM's offerings are not meant to be and never will be general-purpose offerings like ChatGPT. A system based on watsonx is fed limited data sets by design (just what is approved by the owner per security and other requirements), and its "learnings" (what it can report back, for lack of a better word) are strictly controlled by the owning organization to meet its goals and regulations.
So...a watsonx implementation may be trained to spot inconsistencies in credit card authorizations, but it will never have the ability to draw pictures, write poetry or act as a code monkey for the lazy programmer...it doesn't know how. Because data availability and access is heavily segmented and controlled in our society, don't expect a watsonx working for Visa to work the same as a watsonx that works for American Express or MasterCard. The "brain" might be similar, but what it's taught won't be.
Will this be valuable to businesses? I imagine so. Can be compared to other offerings, especially in a sales context? It's way too soon to say. Check back on this offering in 20 years, if IBM is still around by then, and look back on its performance record to see how it's doing.
CP4D offers 9 of the 11 services WatsonX offers. Since CP4D (many times referred as the future of Data and AI) did not take off, IBM added some bells and whistles and branded it as WatsonX.
"@bfaa+1qrd9Ph9 https://dataplatform.cloud.ibm.com/docs/content/wsj/getting-started/compare-platforms.html?context=cpdaas"
In typical IBM fashion, it's not that simple. watsonx builds on things that are in CP4D but most of the GenAI stuff in watsonx is not in CP4D (at least so far). Only IBM has to write a FAQ on how to distinguish between their two highly overlapping products:
https://dataplatform.cloud.ibm.com/docs/content/wsj/getting-started/compare-platforms.html?context=cpdaas
Watsonx.data is essentially cloudpak for data ... so they kinda do.
??? watsonx and Cloud Paks don't really have much to do with each other.
WatsonX is mostly a simple rebranding of another fiasco called cloud paks.
Agree with you... Watsonx should really be called WatConx.
Statement from the office hours was “show me a mode that’s faster at writing Python, Java, cobal, and ansible code”. Real high bar there….
If watsonx is to be anything other than a "me too" offering, then IBM needs some success stories. Right now the public hears all the stories about how good or how bad Siri, Alexa, Google Assistant and ChatGPT are. Each one of them has their strengths and weaknesses, but they are known entities and you can guess what's going on. IBM doesn't have that. Supposedly, watsonx is being targeted to enterprises as a special-purpose product specifically for business use. That's fair enough, but it would be nice to know what watsonx is actually doing for those businesses. Is it helping them to work more efficiently? Does it save time or money? Are they doing something better with watsonx that they weren't doing before? Nobody knows.
Someone out there has to try watson code assistant for z and report the results here. That product has to go down as the most radically unready for prime time product that the company has ever foisted on the public. Arvind is very proud of the fact that he forced the team to deliver it in six months because he is a brilliant leader. The downside to delivering something that promises this much in six months is that what you deliver has not actual chance of doing what IBM claims it can do. Not only that, it probably can't really be fixed. It probably shouldn't be fixed anyway since this is not really an exemplary use of AI. (There are much more promising alternatives.)
watsonx is a scam
Nobody is buying that sh-t
More importantly than investing...You thinking of buying?
AI is such a new field of study that it will take years before everything shakes out and the major players can be determined. Asking who the market players would be like asking about the car makers in the 20th century, or for that matter the computer makers since the 1980s...almost all of them went out of business, and the ones who survived were not the ones that were voted "most likely to succeed" in high school.
For what it's worth...IBM claims 42 clients on their watsonx web page. I dunno if they are real clients, or if they just experimenting for now...I suspect the latter. For a well-heeled company, a few thousand bucks a month for IBM's service is cheap, and watsonx may turn out to be really good...who knows? What's the price for people in the peanut gallery? I'm not enterprise-grade, but I'd be willing to try out watsonx at a competitive price. (Something competitive with peanut-gallery prices from Google, Microsoft or OpenAI.)
You thinking of investing?