BCG Report: IT Firms Lack AI Talent to Capture $200B Opportunity
(a) BCG evaluated 50,000 engineer profiles (b) Only 10% doing Applied AI work at IT firms (c) Urgent need to build Applied AI workforce (d) INSIGHTS for IT engineers:.
Market Is Ready; IT Firms Aren’t
a. Boston Consulting Group (BCG) has published an extensive report on how IT companies can build Applied AI workforce from the people they already employ.
b. Opportunity: BCG estimates in the next 4 years AI will add $400B of new revenue streams for IT services globally, but wipe out $200B of old revenue streams. So, the net gain is $400-$200 = $200B.
c. Problem: IT firms claim to offer “AI-first” services, but in reality they have not built an AI delivery model. These companies only have a fraction of talent needed to capture the vast opportunity.
d. IT companies tag their workforce as “AI engineers,” but in reality only 10-15% do Applied AI work. The rest merely use AI to do old work faster. At AI-native firms (“tech companies”) 90-95% do real Applied AI work.
Infused AI vs. Applied AI
a. Infused AI: It uses AI tools to do existing work faster. An IT engineer’s task is still fundamentally software development, which is now done faster with AI tools.
b. Applied AI: It involves building and operating AI agents that do the work. Examples: enterprise AI agents, agentic workflow automation, production AI platforms, and AI systems for end-to-end business processes.
c. Infused AI eliminates the need of humans for repetitive tasks. If previously 3 engineers worked for 8 hours to develop and test a feature, now only 1 engineer can do the same task. So, two-third of the workforce is gone.
d. Applied AI engineers create new sources of growth & revenue for clients by building and operating AI agents: new services, new products, new customer problems that can be solved economically, and new ways to increase productivity.
e. IT companies lack the capacity in Applied AI. BCG says time is running out, and market opportunity will be grabbed by firms that move fast to build massive “Applied AI” talent.
Applied AI Job Roles
BCG has identified 40 distinct Applied AI roles. Four categories, which IT engineers and IT firms should target are:
a. Forward-Deployed Engineer (FDE): Works closely with the client to design & deliver a working solution.
b. AI Engineer/AI Architect: Builds and scales enterprise AI and agentic systems.
c. Semantic Engineer: Develops data representations, vector embeddings, and knowledge structures that AI agents can use.
d. AI Ops Engineer: Monitors AI systems for performance, accuracy, drift, and bias.
In AI-native firms, 45% of the workforce comprises forward-deployed engineers (FDEs). In traditional IT firms, the share is less than 5%.
5 Questions IT Firms Must Answer
BCG wants IT companies to answer these five questions right now:
a. Has the company selected the AI use cases (agentic systems) it intends to build to solve client problems? Otherwise it may train people in general AI skills without knowing where those skills will be used.
b. Is there consistent role architecture across the company? A common architecture makes job roles, skill requirements, and career paths easier to compare across teams.
c. Has the company made deliberate hiring and training decisions for all 40 Applied AI roles? A blanket hiring and training policy won’t sustain for long.
d. Has the company assessed its existing workforce properly? Accurate assessment prevents the company from overestimating or underestimating its current capabilities.
e. Can the company’s pay and career structures survive competition from AI-native firms? Otherwise the company will lose existing talent or fail to attract top talent.
Endpiece: The Clock Is Ticking
BCG report’s final message is that building AI capability takes time. Winners will be those IT firms that begin early and enable their employees to develop skills and experience before the opportunity matures toward 2030.
Andrej Karpathy described a second brain as a compiler, not a filing cabinet. this diagram shows the architecture behind the idea.
most people treat a second brain like a drawer. notes go in, you search when you need one, and the pile keeps growing without ever getting smarter.
a compiled wiki works differently. raw holds your sources: unstructured, immutable, ground truth. wiki is where the model turns them into structured, linked, evergreen pages. you read it, the model writes it. output is where finished work lands, built from compiled knowledge and not from memory.
at the center sits CLAUDE.md. who you are, what you want, what you're working on. the model reads it before every session, so you never explain yourself twice.
the loop closes when every new source gets compiled back in. five automations keep it running: ingest captures, write drafts, manage ties decisions to context, review reflects, maintain prunes and fixes.
the system gets better with every cycle. search finds what you stored. compiling builds what you understand.
save this, and read the full build in the article below. build the compiler, not the cabinet.
What tasks should AI be trusted with, and when should humans remain involved? Designing those decision rights is one of the defining challenges of the AI era. Researchers at the MIT Center for Information Systems Research have developed a framework to help.
Learn more: https://t.co/NbTi0Vvd7Z
There's a lot of bullshit advice on AI transformation from people who've never made this work in a regulated company.
Here are the 8 steps behind our builds, and what goes wrong when you skip them:
Most ML engineers optimize for accuracy. They should optimize for what happens when the model is wrong. A 95% accurate system that fails silently on 5% of cases is worse than a 75% system that flags its uncertainty.
Today, I shared with the OpenAI team that I have decided to leave my full-time role at OpenAI and transition to being a part-time advisor.
Three months ago, I had to go on medical leave after a severe exacerbation of a chronic illness I’ve lived with for seven years. During that time, it became clear that the road to recovery would be much longer and more complex than I had anticipated—and that I needed to focus on it fully.
When I went on leave, many people told me I was courageous for prioritizing my health. The truth is that I am only making this decision now because I failed to make it many times before.
Over the years, doctors, friends, colleagues, and loved ones encouraged me to slow down. Two years after I got sick, Facebook offered me the opportunity to take a full year of medical leave. I didn’t even pause to consider it. I immediately said no. At the time, Zuck told me I should play the long game. I wish I had listened.
Looking back, I realize that a lot of what made me successful also made this decision incredibly difficult.
I grew up believing that opportunities were precious and that when they appeared, you grabbed them with both hands. That mindset carried me from a small town in southern France to opportunities I never could have imagined. By the time I turned 40, I had already gotten to do more than I’d ever dreamed possible as a kid growing up in Sète.
I love building. My work has always given me a deep sense of purpose. OpenAI in particular felt like a role that my entire career had been building toward, which made this decision even harder.
But what I’m learning now is that grit and endurance are not the only skills required to have impact over decades. Sometimes the harder thing is to stop, listen, and trust that taking care of yourself today makes it possible to contribute for much longer tomorrow.
This experience has also strengthened my conviction about why this work matters.
It has been a jarring experience to spend my days helping build the future while simultaneously navigating a disabling disease that still has no cure.
Over the last seven years, I’ve spent countless hours in doctors’ offices, dealing with symptoms, treatments, insurance, uncertainty, and all the invisible work that comes with being a patient. Like millions of others living with chronic illness, I’ve experienced firsthand how difficult healthcare can be to navigate, even when you have every possible advantage.
More than ever, I believe that some of the most important opportunities for AI lie in helping people solve real problems in their daily lives: their health, their finances, their time and the everyday burdens that shape human experience.
In particular, curing disease is the most important thing AI could accomplish. I’m excited to continue working towards cures through OpenAI but also through my work with @ChronicleBioAI and @CODA_research.
I’m deeply grateful to @sama, @gdb and the OpenAI board for their support during this time and for offering a way for me to continue contributing to the mission without sacrificing my chances of recovery. I’m also so thankful to my team and the many extraordinary colleagues I’ve had the privilege to build alongside.
For now, my focus is recovery. But my belief in the potential of technology to solve deeply human problems has never been stronger.
Apple's newest patent turns your AirPods into an EEG device that also reads muscle movement, eye movement, and heart rate.
Neuralink has performed four human brain implants. Apple has hundreds of millions of AirPods in active use.
Steven Hotelling is on the inventor list for patent US20230225659A1. He's the engineer who productized multi-touch for the original iPhone. Apple puts Hotelling on things it plans to ship.
The design packs 17 electrodes around a single ear tip. Ear-EEG has existed in academic research since 2011, and startups like NextSense (spun out of Alphabet X), Neurable, and IDUN already ship consumer ear-EEG products today.
Apple's innovation is dynamic electrode selection.
Here's the constraint that killed ear-EEG as a consumer product for 15 years. Every human ear canal is shaped differently. A fixed electrode placement that reads clean EEG from one person reads noise from the next. Your own ear canal changes shape across a day based on temperature, jaw position, and how the bud seated when you put it in. Every prior ear-EEG product required either custom molds or tolerance for garbage data.
Apple's patent uses more electrodes than are ever needed simultaneously and runs an AI model that scores each one in real time on impedance, noise level, and skin contact quality. It picks the best subset for this person, in this ear, right now. Reference and active electrodes get reassigned on the fly. A weighted algorithm combines the surviving signals into one optimized waveform.
Competitors tried to solve this with better electrodes. Apple solved it with redundancy and silicon.
The roadmap is already visible. AirPods Pro 3 shipped a photoplethysmograph in the ear tip for heart rate. In November 2025, Apple Research published PARS, a self-supervised model that learns EEG patterns from unlabeled data and sidesteps the regulatory bottleneck of annotated clinical datasets. The PPG sensor is the dress rehearsal. The EEG sensor is the feature.
Sleep staging. Seizure detection. Stress classification. Focus state. The clinical applications get through the FDA. The interface applications change computing.
Neuralink needs a surgeon. Apple needs a firmware update.
Most product managers believe they manage the product.
What they actually manage is the experience people have with the product.
One can quickly shift from believing they're doing well to having trouble sleeping once they see this.
PMs don’t just ship features. They kill them. Shipping isn’t the job. Shipping the right product is. A great PM doesn’t fall in love with the roadmap. They fall in love with the problem and have the guts to say: This isn’t solving it. This adds complexity. This doesn’t matter. Every feature, setting, UI, element should fight to exist.
At Nest, we had one rule:
If you can’t explain why it matters, it doesn’t ship.
You had to tell us the why.
The reason a real person would care.
That one rule killed dozens of features.
JPMorgan is not experimenting with AI. They’ve operationalized it.
- 450+ internal tools.
- 200,000 employees use them daily across engineering, finance, and customer ops.
Here is a breakdown of their full AI stack: https://t.co/TTLE5EqJnH