Thread regarding SAS Institute layoffs

Dinsmore's Assessment

"It's 2017...

Your SAS Account Executive wants a meeting. She wants to tell you all about SAS Viya, SAS’s “open, cloud-ready, next-generation high-performance analytics and visualization architecture.” SAS launched Viya last year, and it’s a dud.

Viya is SAS’s “answer” to Apache Spark. Like Spark, it’s a distributed in-memory computing system; SAS claims, without evidence, that Viya is faster than Spark. It’s certainly a lot more expensive than Spark. The SAS applications that run on Viya have graphical user interfaces, so it’s easier to use than Spark. But nobody wants to hand petabyte-scale applications to low-IQ users who can’t write code.

You still have numerous SAS users. Most of them work on their desktops with plain old SAS: Base, STAT, and Graph, with Enterprise Guide. They will continue using SAS unless you tear it from their cold, dead hands, which nobody will do as long as their department budget covers the license fees.

SAS is secure in your company until the existing users die or retire. However, the last thing you want to do is expand your SAS footprint or invest in a “next generation” architecture that will lock you in for another 20 years."

https://thomaswdinsmore.substack.com/p/a-short-history-of-automl-part-eight


by
| 3418 views | | 29 replies (last November 15) | Reply
Post ID: @OP+1k70p58aq

29 replies (most recent on top)

"If not a “blunder”, Viya was at least a strategic mistake.

However:if industry trends were not clear at that time, they certainly are undeniable now, ten years later."

It has been said before and is worth repeating. Viya was the inflection point that marked the beginning of the end for SAS. A submarine can descend only so deep intact before it implodes.

by
| | Reply
Post ID: @5qq+1k70p58aq

@jn “CAS and related Viya components were well aligned with the SAS leadership and culture of 2010-2015.”

True. It’s also true that Spark was new at that time. But R had been around much longer. If not a “blunder”, Viya was at least a strategic mistake.

However:if industry trends were not clear at that time, they certainly are undeniable now, ten years later.

Yet SAS has not changed its strategy. It maintains two incompatible platforms, each competing with Open Source.

Good luck to all.

by
| | Reply
Post ID: @1bt+1k70p58aq

@135 that’s why I come here. I always learn something! :)

by
| | Reply
Post ID: @143+1k70p58aq

A company based on a cult of personality supported by magic gnomes. What’s not to love?

Stack it high and deep while you can.

by
| | Reply
Post ID: @132+1k70p58aq

"Don’t need magic gnomes. Already have long term career prosperity at SAS."

I think the point is looking prospectively, not retrospectively. SAS was very good for the long-term career prosperity of thousands of employees, me included, and I'll forever be grateful. However, it does appear that SAS is now facing (and has for a number of years) significant headwinds for this to continue to be the case.

by
| | Reply
Post ID: @12x+1k70p58aq

@108 @qs wrote that denying Dinsmore will not bring long-term career prosperity.

That’s because SAS been shrinking for many years, with no prospect of improvement. As Dinsmore stated, Viya has not halted the decline.

Nobody made any assumptions about your personal finances. If you have already achieved prosperity at SAS, good for you. But that is not looking long-term.

by
| | Reply
Post ID: @12k+1k70p58aq

“ In a shrinking company, no one has long-term career prosperity.”

Ok you are clearly wise in the affairs of every single person. Well done.

With that kind of insight you must be prosperous indeed with whatever path other than SAS you chose. But I can’t speak to that because I don’t know who you are and don’t have the same mentalist skills you do.

Meanwhile I’m going to go look at my bank accounts and despair. Wish I had done something more prosperous during my career.

by
| | Reply
Post ID: @108+1k70p58aq

@101 SAS revenues have declined ~2% against inflation for over a decade.


In consequence, SAS has shrunk headcount by ~2% every year, over that same period of time.

In a shrinking company, no one has long-term career prosperity.

by
| | Reply
Post ID: @102+1k70p58aq

@zt+1k70p58aq Not kool-aid. Fact.

Spoken as the only person on this site who can actually speak for myself.

If you think you can blindly say KoolAid without knowing the first thing about me that speaks more about you than me.

And I’m far from the only person who has had and continues to have long term career prosperity.

Big a-s batch of kook aid.

by
| | Reply
Post ID: @101+1k70p58aq

“Already have long term career prosperity at SAS.”

Damn, that is strong Kool-Aid.

by
| | Reply
Post ID: @zt+1k70p58aq

“ while magic gnomes pay your salary will not bring you long-term career prosperity.”

Don’t need magic gnomes. Already have long term career prosperity at SAS.

And ample opportunity to go elsewhere despite y’all thinking everyone at SAS must be a mo--n. I choose to stay.

You do you.

by
| | Reply
Post ID: @zn+1k70p58aq

Ask for a list of customers and the annual revenue amounts for your product. See if you get a solid answer. (You won’t)

Deny Dinsmore all you like, but playing Montessori software games while magic gnomes pay your salary will not bring you long-term career prosperity.

The cognitive dissonance is real.

by
| | Reply
Post ID: @qs+1k70p58aq

You seem unaware that Dinsmore’s series of articles is a retrospective

More like retrospeculative!

by
| | Reply
Post ID: @qr+1k70p58aq

“ My denial and delusions will keep me safe, right?! Right?! RIGHT?!?!?!”

And ever increasing capitalizations coupled with increased punctuation make you seem smart. right? Right?! RIGHT?!?!!!?

by
| | Reply
Post ID: @qj+1k70p58aq

Best sleep aid ever was OS' seemingly never ending eye glazing delivery of Viya rollout at Vegas.

That was the beginning of the end of SAS' greatness. Thomas D nailed it.

by
| | Reply
Post ID: @pq+1k70p58aq

@k2 Once again, please permit me to help.

adjective
ret·​ro·​spec·​tive | \ ˌre-trə-ˈspek-tiv \
Definition (Entry 1 of 2)
1 a (1) : of, relating to, or given to retrospection
(2) : based on memory
//a retrospective report

noun
ret·​ro·​spec·​tion | \ ˌre-trə-ˈspek-shən \
Definition
: the act or process or an instance of surveying the past

by
| | Reply
Post ID: @ne+1k70p58aq

Wait?! The house is on fire?! That can’t be! I’m going to close my eyes and hide under the bed while chanting “the fire isn’t real, the fire isn’t real, the fire isn’t real”. My denial and delusions will keep me safe, right?! Right?! RIGHT?!?!?!

by
| | Reply
Post ID: @my+1k70p58aq

You seem unaware that Dinsmore’s series of articles is a retrospective.

And, as such, it should be forward looking and predictive!

by
| | Reply
Post ID: @k2+1k70p58aq

@ew You seem unaware that Dinsmore’s series of articles is a retrospective.

by
| | Reply
Post ID: @jw+1k70p58aq

… Except for the Hadoop/related ecosystem and R

by
| | Reply
Post ID: @jp+1k70p58aq

From AI search: “Apache Spark's initial release was on May 26, 2014. While development began earlier at UC Berkeley in 2009 and it was open-sourced in 2010, the first official release under the Apache Software Foundation was in 2014.”

In the same timeframe: CAS, supporting micro services and updated V* products, eventually branded as Viya were either in design or already under development.

Except for Hadoop, Open Source data management and analytics were only beginning to gain adoption. SAS was still riding tall as the undisputed leader in advanced analytics with considerable internal resistance to any serious participation in the open source ecosystem — especially as a contributor.

The SAS customer base was growing and knowing SAS software (across a broad range of specific career skills) was considered very valuable in many industry sectors.

For these and many other reasons (covered extensively on earlier threads), CAS and related Viya components were well aligned with the SAS leadership and culture of 2010-2015.

That said, Dinsmore is a good pundit and his analysis of historical product blunders is often correct.

by
| | Reply
Post ID: @jn+1k70p58aq

That was what was so foolish about building Viya. Why build something that is freely available in the form of Spark?

SAS should have built GUI applications on top of Spark...industry solutions that SAS has expertise and heritage in, like credit risk, fraud, etc. That could have worked.

by
| | Reply
Post ID: @hx+1k70p58aq

Dinsmore makes clear that, even when it launched, Viya competed against Spark. That’s free, so Viya can’t compete on price.

But some number of users want GUIs. They’ll buy Viya, if the price is affordable.

Some number of users are happy with SAS9, and will never switch, to Viya or to anything else.

The concerning statement is that, at Dinsmore’s installation, management did not want to expand their SAS footprint at all; it was being kept until existing users died or retired.

by
| | Reply
Post ID: @gm+1k70p58aq

@e0+1k70p58aq Impressive. An article about state of ML as of 2017 published in 2025.

by
| | Reply
Post ID: @ew+1k70p58aq

How dare you call or infer that my baby is ugly?! Damn you and your real world anecdotes!

by
| | Reply
Post ID: @e1+1k70p58aq

@d8 The blog was published one week ago.

by
| | Reply
Post ID: @e0+1k70p58aq

Hey let’s post last updated 2017 blog series one more time.

by
| | Reply
Post ID: @d8+1k70p58aq

The weakest link defines the strength of a chain.

by
| | Reply
Post ID: @cg+1k70p58aq

Read through each installment and you’ll notice a pattern— increased reliance on other tools.

by
| | Reply
Post ID: @as+1k70p58aq

Post a reply

: