Thread regarding SAS Institute layoffs

This summer will see 14 new interns in the Art Department, which is safe.

Perhaps some will resort to the ‘SAS Beg’ tactic - it was successful in the past.

“the real question for SAS is if they have any comprehension of how to deal with ‘true’ AI…”

For a company cutting costs, the best move is to replace some of its junior people with AIs supervised by senior people.

The senior people have enough experience to correct the AI’s mistakes. The company only needs to retain and train a few juniors, to replace the seniors when they retire.

SAS has been doing the opposite, laying off / buying out / pushing out seniors, while replacing them with juniors. That is cost-cutting on a basis of simple arithmetic, eliminating the highest salaries.

AI has made that sort of cost-cutting obsolete. Going forward, SAS’s cost-cutting policies should adjust to the new reality.

SAS will continue to emphasize regulated industries such as financial services, healthcare and government.

The U.S. government does not need software assistance in its fraud, waste, and abuse. It's currently doing just fine, thank you very much. (And thank you very much for your attention to this matter).

While there is plenty of speculation on the impact of AI on hiring, especially for junior people, the real question for SAS is if they have any comprehension of how to deal with “true” AI and compete. For that reason, SAS will continue to emphasize regulated industries such as financial services, healthcare and government. Opportunities outside these markets will get swallowed up by competitors.

@3q3 Stanford and Harvard studies confirm that AI is reducing hiring of juniors — at most companies. SAS marches to its own drum.

https://fortune.com/2025/08/26/stanford-ai-entry-level-jobs-gen-z-erik-brynjolfsson/

https://www.reddit.com/r/InformatikKarriere/comments/1nj4r8r/harvardstudie_es_ist_vorbei_für_berufseinsteiger/?tl=en

But…but…but…. We’ve caught the ‘SAS Bug’ (the SAS Beg) after our internship. Surely, you will take pity upon us and hire us?

(I am not making this up. They were all hired after this ploy).

Time to sell the toney Cary properties, build a tiny house on some cheap land an hour from nowhere and invest whatever you can in the hyperscalers so as to reap the next three years or so of gains.

@3j7 Your logic is flawless.

For example, an IT department can replace a team of one Senior and three Junior developers with one Senior, one Junior, and an AI.

The cost of the AI is more than offset by the saving of two salaries, two benefit packages, and two software licenses.

The math is so simple, they’ll all figure this out.

And the license fees will flow away from the vendors of proprietary languages, and toward the hyperscalers.

@3c2

also, leaving aside questions of design and architecture, LLMs can generate python well enough. the need for particular languages themselves, especially proprietary, or armies of humans who can write instructions in those languages, is further reduced. ergo, demand for licenses for software running those languages, especially proprietary and expensive, also reduces. further, agentic AI likely means foundational LLMs or adjacent businesses are now the platforms to run that code and applications that are "vibe coded" or "agent coded". so demand for cloud platforms for humans to write open source or proprietary code then deploy applications is likely reduced. the entire value chain is changed, except that hyperscalers are likely still in the best positions to adapt to the rapid changes.

SAS has never paid attention to design, and never valued it.

SAS products are the result of “Feature Factories”: a collection of features, with no coherent design. When this does not delight the customers, they go elsewhere.

Apple’s products are a result of design. Apple engineers resent that designers are in charge. But Apple products sell.

SAS hires from NC State, which has a fine School of Design. But the best graduates of that school can make more money at Apple or Google.

And if they go to SAS, SAS will not use their expertise.

Design - A pretentious field where practitioners operate under the delusion that their efforts can save the world.

The Art Department, which is safe.

Says the person that continues to hold a grudge since failing third grade arts and crafts.

@3d0

The Art Department, which is safe. :/

@3cy product designers :)

Prediction: over the next 3 to 5 years, AI technology will replace the need for a significant number of data management and analytics tools — especially those requiring annual muli-million $ renewals because they are obviously targets.

Innovative, lithe younger companies moving much faster than SAS are already producing Agentic AI suites that will eat into SAS’ vertical product revenues.

SAS could respond with MAJOR management changes (exodus of old guard) immediately, termination of waste/bloat and reinvestment in hiring of top product designers and architects to build a new AI centric approach that jettisons the traditional SAS language and infrastructure.

The probability of this approaches zero under current ownership.

Let’s be honest in the real world: There is no Open Source per se, there is Python. It is the universal programming language for data science and does much more. The SAS programming language has great value, but it is almost impossible to convince the public it is better than Python. I would say R faces the same challenge.

SAS, as a company, needs applications that makes LLM customization, like RAG and LORA easier. This will open to the door to “true” Agentic AI and tools like Intelligent Decisioning.

You integrate it with Prego Studio. - “It’s in there!”

Most of you say you would do exactly what SAS did: support interoperability with R, Python, etc., but not use Open Source as a required foundation.

I don’t see any way to do that either, without cannibalizing revenues.

The Innovator's Dilemma… describes how large incumbent companies lose market share by listening to their customers and providing what appears to be the highest-value products, but new companies that serve low-value customers with poorly developed technology can improve that technology incrementally until it is good enough to quickly take market share from established business.

https://en.wikipedia.org/wiki/The_Innovator's_Dilemma

-- @OP

It's not an either-or scenario. Open source and homegrown SAS technology can coexist in SAS Platforms. Our offerings will then bring the best of both worlds, in a way that's transparent to users.

I had led integration of multiple open source packages into SAS offerings under the hood before I left SAS. Such hybrid modules we built made their way into multiple SAS products and brought lots of revenue.

How would you integrate SAS with Open Source?

What do you mean? Open source is already integrated with SAS and Viya.

For example, you can already run Python from the DATA step. You can run R code. What more do you want?

@OP

There are plenty of smart people at SAS who have an answer to your question, and even more smart people who once worked for SAS but have since left the company who had an answer to your question. But the better question is: why isn't this already the case (that SAS was successfully "integrated" with Open Source)? If you understand the answer to the better question, you'll understand why answering yours thoughtfully is mental masturbation.

@OP Technically we use large amounts of Open Source.