Managing Editor’s Notice: Our colleague Larry Benedict, founding father of The Opportunistic Dealer, doesn’t chase moonshots which have already left orbit.
As an alternative, he makes use of ONE ticker for a shot at extra features in a ONE day than the S&P 500 usually delivers in a complete YEAR. Again and again. And he does it with out ever shopping for, promoting, or holding a single inventory.
Anybody can study to do that. That’s why on Thursday, Might 7, at 8 p.m. ET, he’s exhibiting you the best way… together with the ONE ticker that makes all of the distinction.
Click here to save your spot automatically and find out more.
At face worth, the narrative is logical.
Synthetic intelligence (AI) is enabling the most important productiveness enhance in historical past.
It has already confirmed to be exceptionally good at software program programming and operations.
AI can carry out the identical quantity of labor as a staff of people, and all it wants is compute and electrical energy.
Due to this fact, software program builders and programmers will lose their jobs en masse.
That is exactly the narrative that decels and naysayers have been pushing in an effort to sluggish issues down and/or develop heavy laws for AI.
Besides there’s one main drawback.
It’s not true.
The alternative is occurring.
Early False Flags
For a brief time period from late 2023 to summer time 2025, it did seem like job alternatives for software program engineers have been on the decline, as proven within the chart beneath.
Supply: Citadel Securities
However this was only a false flag.
Three key issues have been taking place in high-tech throughout that window:
- Tech corporations have been (and nonetheless are) aggressively deploying AI of their organizations.
- Tech corporations have been attempting to determine learn how to finest make the most of and leverage AI for software program improvement.
- Tech corporations have been utilizing the expertise as an excuse to reorganize and clear out any bloat and underperformers from their organizations.
Let’s take a look at an ideal instance of this with Salesforce (CRM).
In September of final yr, not way back, Salesforce CEO Marc Benioff introduced that he was reducing 4,000 of the corporate’s buyer assist roles.
This represented about 5.2% of the huge software program firm’s worldwide headcount.
In his phrases, when referring to the influence of AI on his buyer assist group: “I’ve diminished it from 9,000 heads to about 5,000, as a result of I would like much less heads.”
Supply: CNBC
Once more, at face worth, it seems to be like AI is having a direct influence on the discount of jobs on the software program firm.
However then, let’s take a look at this, a publish on X from Marc Benioff from April 24 this yr:
Supply: Marc Benioff
Months later, Benioff pronounces that he’s hiring 1,000 new grads and interns to construct with synthetic intelligence.
Not one of the media has acknowledged the apparent.
An Straightforward Off-Ramp
Each two or three years, generally extra continuously, tech corporations have a discount in drive (RIF) of 5-10% of their world workforce.
Corporations typically develop into bloated or lazy and want to scrub out the components of the group that merely aren’t performing nicely.
In excessive tech, metrics are extraordinarily vital.
You’ll typically hear a phrase like one headcount is one headcount, so we have to have one of the best individual for the job. That form of mindset is important to compete and develop in excessive tech.
So, these sorts of bulletins are used to scale back the underside 5-10% of the performers and infrequently to have an effect on some form of restructuring.
Once I labored within the semiconductor trade, this sort of RIF would occur a minimum of yearly. It might sound brutal, however the necessity to innovate and enhance within the semiconductor trade is incessant.
Conducting a big RIF additionally gives the corporate with authorized cowl to eradicate troublesome workers – recognized unhealthy actors, who may be a threat by trying to sue the corporate for some form of fabricated mistreatment.
So, utilizing AI gives a fantastic excuse to scrub up a corporation and start the method of hiring higher individuals.
One thing very totally different is occurring now in excessive tech, and it’d really feel counterintuitive.
AI will not be leading to much less work for software program engineers.
AI is leading to extra work for software program engineers.
From my very own community, I’m listening to that software program engineers are busier than ever.
Agentic and generative AI is enabling them to be extra productive.
Reasonably than working much less, they’re working extra. They’re busier than ever earlier than.
They’re leveraging the newfound expertise to do issues in days that will have taken them a yr up to now. And they’re ridiculously excited concerning the transformation that’s underway, so they’re reengaging and leaning in.
Why is that this taking place? Shouldn’t it’s the alternative?
No, it really makes lots of sense.
Management, Don’t Vibe
A analysis paper was not too long ago revealed about precisely this subject, which does a fantastic job at explaining why that is true.
Professional Software Developers Don’t Vibe, The Control: AI Agent Use for Coding in 2025 labored with and surveyed 112 skilled software program builders to higher perceive the influence of the expertise on their jobs.
The analysis discovered that agentic and generative AI have been nice at supporting sure duties and never most well-liked for a variety of different duties.
For instance, the expertise was discovered to be extremely helpful to software program engineers within the sorts of duties listed beneath.
And alternatively, it was discovered to even be unsuitable for a variety of duties, that are proven beneath.
To summarize what was discovered, AI was not very nicely suited to simply one-shotting code, architecting a software program program, performing complicated duties, or making use of enterprise logic to software program improvement.
Software program builders have been nonetheless wanted to carry out the planning and the software program structure after which to design the prompts for the AI for very particular duties that wanted to be carried out. Understanding learn how to “constrain” or restrict the AI in a given job is a important human-level job operate.
Additionally they discovered the necessity to assessment each change made by the AI to grasp the variations within the software program.
It’s one factor to generate the code, nevertheless it’s one other to check it and be sure that it’s prepared for placing into manufacturing and is nice sufficient as an enterprise-grade software program bundle.
The truth is that to make use of agentic and generative AI successfully, it needs to be tightly monitored and managed.
When it’s not, it ends in high quality points and software program slop that may create liabilities for a corporation.
Jevons Paradox Unleashed
That is exactly why there’s a rise in demand for software program engineers.
Corporations want each new graduates and skilled software program engineers to correctly harness the facility of agentic and generative AI.
And each tech firm is aware of that in the event that they don’t lean in and reap the benefits of this productiveness boosting expertise, they are going to be left behind by the competitors.
The very fact is, AI makes the event of high-quality software program cheaper and extra environment friendly.
The result’s that corporations at the moment are in a position to construct much more software program and extra worth, enabling extra progress.
It’s a traditional Jevons Paradox: When one thing helpful turns into cheaper and extra economically accessible, adoption and utilization soars.
That’s precisely what’s taking place now, and it’s exactly why the software program trade isn’t lifeless in any respect.
Software program engineering jobs aren’t going away.
The trade is restructuring itself and adapting to the mass employment of, and deployment of, agentic and generative AI.
Jeff