Latest / DX Today | No-Hype Podcast & News About AI & DX / The Fed Says AI Is Quietly Shrinking Hiring in 5 of 12 Districts

Transcript

The Fed Says AI Is Quietly Shrinking Hiring in 5 of 12 Districts · DX Today | No-Hype Podcast & News About AI & DX

• 0:00
  SPEAKER_00: Welcome to the DX Today Podcast, your daily deep dive into the AI ecosystem.
• 0:05
  SPEAKER_00: I'm Brian, and joining me as always is Jessica.
• 0:08
  SPEAKER_01: Thanks, Brian.
• 0:09
  SPEAKER_01: Today we are digging into a document most people have never read, the Federal Reserve Beige Book.
• 0:14
  SPEAKER_01: Because this week it quietly said something about AI and jobs that nobody at the Fed had put on paper this way before.
• 0:21
  SPEAKER_00: I will admit the beige book is not exactly beach reading.
• 0:24
  SPEAKER_00: So before we get to the AI part, give people the 30-second version of what this thing actually is and why anybody should care.
• 0:31
  SPEAKER_01: Sure.
• 0:32
  SPEAKER_01: Eight times a year, each of the twelve regional Federal Reserve banks talks to business owners, bankers, staffing firms, and manufacturers in its district, and the Fed publishes what they heard.
• 0:43
  SPEAKER_01: It is anecdotes, not statistics, but it is anecdotes collected at enormous scale by people whose job is to listen.
• 0:50
  SPEAKER_00: So it is basically the Fed walking around the shop floor and asking the foreman how things are going, rather than waiting three months for the official spreadsheet to tell them what already happened.
• 1:00
  SPEAKER_01: Exactly that.
• 1:01
  SPEAKER_01: This edition came out on Tuesday, October 14th.
• 1:04
  SPEAKER_01: It was compiled by the Philadelphia Fed, and the information was collected on or before October 5th.
• 1:10
  SPEAKER_01: And in the national summary, there is one sentence that has been bouncing around every economics desk since.
• 1:16
  SPEAKER_00: All right, read me the sentence, because I suspect the exact wording matters a lot here, and I would rather hear what the Fed actually wrote than somebody's headline about it.
• 1:25
  SPEAKER_01: Here it is.
• 1:26
  SPEAKER_01: Five districts reported that some firms had either reduced staffing levels through attrition, delayed hiring, or anticipated less job creation after increasing productivity by investing in AI.
• 1:38
  SPEAKER_01: Five out of twelve, Brian.
• 1:39
  SPEAKER_01: That is not one quirky contact in one city.
• 1:42
  SPEAKER_00: Okay, but I want to slow down on the verbs before we panic, because reduced through attrition, delayed hiring, and anticipated less job creation are three very different things from a wave of pink slips landing on desks.
• 1:55
  SPEAKER_01: You are right.
• 1:56
  SPEAKER_01: And that distinction is actually the whole story.
• 1:59
  SPEAKER_01: Nobody in this report is describing mass layoffs.
• 2:02
  SPEAKER_01: What they are describing is a door that is slowly closing rather than a door being slammed.
• 2:07
  SPEAKER_01: And those two things feel very different depending on which side of the door you are standing on.
• 2:11
  SPEAKER_00: Explain the door metaphor a little more for the people doing the work, because from where I sit, attrition sounds like the gentlest possible version of this.
• 2:18
  SPEAKER_00: Nobody gets fired, and the company just does not backfill.
• 2:22
  SPEAKER_01: For the people already inside, it is gentle.
• 2:25
  SPEAKER_01: If you have a job, attrition barely touches you.
• 2:28
  SPEAKER_01: But every role that is not backfilled is a job that someone outside never gets offered, and the people outside are disproportionately the 22-year-old graduate looking for a first rung on the ladder.
• 2:38
  SPEAKER_00: That lines up with something I have seen running teams.
• 2:41
  SPEAKER_00: When a company decides not to replace someone, nobody announces it.
• 2:45
  SPEAKER_00: There is no press release, and it never shows up in a layoff tracker.
• 2:49
  SPEAKER_00: So it is the quietest way a labor market can change.
• 2:52
  SPEAKER_01: Which is precisely why the beige book matters here.
• 2:55
  SPEAKER_01: The official data, unemployment, payrolls, jobless claims, are built to catch people losing jobs.
• 3:01
  SPEAKER_01: They are much worse at catching jobs that simply never got created in the first place.
• 3:05
  SPEAKER_01: And this report is picking up that invisible part.
• 3:08
  SPEAKER_00: Let's get specific because I always trust the details more than the summary.
• 3:12
  SPEAKER_00: Which districts actually said what?
• 3:14
  SPEAKER_00: And are we talking about banks, factories, software shops, or something else entirely?
• 3:18
  SPEAKER_01: Chicago had a business services contact who said productivity gains from AI adoption allowed their company to reduce employment through attrition.
• 3:26
  SPEAKER_01: Cleveland had two banking contacts saying automation and AI had reduced staffing needs, and those same contacts complained about the rising cost of software, AI tools, and hardware.
• 3:36
  SPEAKER_00: I love that detail from Cleveland, by the way, because it is the most honest thing in the whole report.
• 3:41
  SPEAKER_00: The banks are saying we need fewer people, and also our software bill is going up, which is exactly the trade every operations manager is quietly making right now.
• 3:51
  SPEAKER_01: Then Dallas.
• 3:52
  SPEAKER_01: According to reporting from Token Post, an accounting contact there said some large firms are holding off on signing new hires because they expect to need fewer entry-level workers, partly because of AI efficiency.
• 4:03
  SPEAKER_01: That one stings because accounting has always been a classic first job.
• 4:07
  SPEAKER_00: And Minneapolis had the line I think is the most useful framing in the entire document, if I remember the coverage correctly, something about AI dampening job creation rather than causing layoffs.
• 4:18
  SPEAKER_00: Right.
• 4:19
  SPEAKER_01: Information technology staffing contacts in Minneapolis said AI was dampening the creation of new jobs, but was not causing layoffs.
• 4:28
  SPEAKER_01: Honestly, if you wanted to summarize the whole report in one sentence, that is the sentence I would pick, because it captures the shape of what is happening.
• 4:35
  SPEAKER_00: Now let me play the skeptic I promised to be, because I do not want us to oversell this.
• 4:40
  SPEAKER_00: Atlanta said something pretty different, didn't they?
• 4:42
  SPEAKER_00: And I think that deserves equal airtime alongside the gloomier districts.
• 4:46
  SPEAKER_01: They did.
• 4:47
  SPEAKER_01: Atlanta said AI implementation broadened across surveyed firms, with more contacts deploying AI tools and automation.
• 4:54
  SPEAKER_01: But most firms did not expect those efforts to lead to significant workforce reductions in the near term.
• 5:00
  SPEAKER_01: So adoption is spreading, but the cuts are not following everywhere.
• 5:03
  SPEAKER_00: And then there is the other side of the ledger where AI is clearly creating work.
• 5:08
  SPEAKER_00: Philadelphia credited manufacturing growth partly to data centers, AI and defense, and New York had a tech consultancy calling client demand for AI data preparation and governance a bright spot.
• 5:20
  SPEAKER_01: Which is the honest picture.
• 5:21
  SPEAKER_01: The money flowing into AI is creating jobs in construction, electrical work, and consulting.
• 5:27
  SPEAKER_01: And that same report says shortages persisted for specialized, technical, and skilled trade workers.
• 5:33
  SPEAKER_01: The problem is the new jobs and the vanishing jobs are not the same people in the same places.
• 5:37
  SPEAKER_00: That is the part I think gets lost.
• 5:40
  SPEAKER_00: Telling a 23-year-old who expected to start as a junior analyst that the economy needs electricians is technically true and practically useless for that person on Monday morning.
• 5:50
  SPEAKER_01: And the timing makes it sharper.
• 5:52
  SPEAKER_01: According to figures summarized by 24-7 Wall Street, unemployment ticked up to 4.2% in September.
• 5:58
  SPEAKER_01: From 4.1, inflation is running around 3.7%.
• 6:01
  SPEAKER_01: And the Fed actually raised rates in September.
• 6:04
  SPEAKER_01: So this is not a booming market absorbing everyone.
• 6:07
  SPEAKER_00: So the backdrop is a labor market that is soft but not collapsing.
• 6:11
  SPEAKER_00: Inflation that is still too hot.
• 6:13
  SPEAKER_00: And now a central bank hearing from its own contacts that a new technology is trimming hiring at the margin.
• 6:20
  SPEAKER_00: That is an uncomfortable combination for anyone setting policy.
• 6:24
  SPEAKER_01: It really is.
• 6:25
  SPEAKER_01: And Fed Governor Lisa Cook spoke right into that discomfort a couple of weeks ago.
• 6:30
  SPEAKER_01: On September 28th, at the opening keynote of Oakland Tech Week, she gave a speech called An Update on AI and the Economy.
• 6:37
  SPEAKER_01: And she was surprisingly candid about the limits of the Fed here.
• 6:40
  SPEAKER_00: Candid how?
• 6:42
  SPEAKER_00: Because central bankers are usually trained to say as little as possible in as many words as possible.
• 6:47
  SPEAKER_00: So I am curious what counts as candid at that level.
• 6:50
  SPEAKER_01: She said the labor market has remained remarkably resilient through the early stages of AI adoption, and that unemployment and layoffs remain low.
• 6:57
  SPEAKER_01: But she also said recent college graduates are having more trouble finding first jobs, which may partly reflect AI handling entry-level tasks.
• 7:05
  SPEAKER_01: And here is the candid part.
• 7:07
  SPEAKER_00: Go on, I am listening.
• 7:08
  SPEAKER_00: And I suspect this is where the conversation gets genuinely uncomfortable for anyone who assumes the Fed can simply fix whatever goes wrong.
• 7:15
  SPEAKER_01: She described a scenario where AI pushes unemployment up, at least temporarily, because of skill mismatches rather than weak demand.
• 7:23
  SPEAKER_01: And she said in that case, the Fed would have limited tools.
• 7:26
  SPEAKER_01: Interest rates cannot retrain an accountant.
• 7:29
  SPEAKER_01: Cutting rates does not teach anyone how to manage an AI workflow.
• 7:32
  SPEAKER_00: That is a remarkably important admission, because the entire public mental model is that when jobs get scarce, the Fed cuts rates and things recover.
• 7:41
  SPEAKER_00: If the problem is structural, that playbook just does not work, and nobody has a second playbook ready.
• 7:47
  SPEAKER_01: And the global institutions are saying the same thing louder.
• 7:51
  SPEAKER_01: This week in Bangkok, at the annual meetings of the International Monetary Fund and the World Bank, the fund's managing director, Kristalina Georgieva, called AI a slow-moving tsunami for the labor market.
• 8:03
  SPEAKER_00: Slow-moving tsunami is a vivid phrase, but I want to know if she backed it up with numbers, because I have heard a lot of dramatic metaphors about AI and jobs over the last few years, and most of them were just vibes.
• 8:14
  SPEAKER_01: She did.
• 8:15
  SPEAKER_01: The fund estimates about 40% of jobs worldwide will be affected, enhanced, downgraded, or eliminated, rising to around 60% in advanced economies.
• 8:26
  SPEAKER_01: And she raised a question that I think every finance ministry should be losing sleep over.
• 8:30
  SPEAKER_01: With less labor and more capital, what do we tax?
• 8:34
  SPEAKER_00: That tax question is fascinating, and it is the second-order effect you always chase.
• 8:39
  SPEAKER_00: Payroll taxes fund a lot of the safety net.
• 8:42
  SPEAKER_00: So if companies grow output while holding headcount flat, the money they save does not automatically flow back into the system that catches displaced workers.
• 8:50
  SPEAKER_01: Exactly.
• 8:51
  SPEAKER_01: And look at where the money is flowing instead.
• 8:53
  SPEAKER_01: Goldman Sachs estimates Amazon, Microsoft, Alphabet, and Meta are collectively on pace for about$757 billion in capital spending this year.
• 9:05
  SPEAKER_01: That is money going into chips, buildings, and electricity, not salaries for junior staff.
• 9:10
  SPEAKER_00: Let me push back a little though, because I think there is a lazy version of this story where every hiring slowdown gets blamed on AI.
• 9:16
  SPEAKER_00: Companies also overhired a few years ago.
• 9:19
  SPEAKER_00: Borrowing costs are high, and blaming the robots is a more flattering explanation for a CEO than admitting weak demand.
• 9:26
  SPEAKER_01: That is a fair point.
• 9:27
  SPEAKER_01: And I will not pretend otherwise.
• 9:29
  SPEAKER_01: 24-7 Wall Street cited a tally where AI was named as a factor in more than 120,000 announced job cuts through September, but announced as a factor is doing a lot of work in that sentence.
• 9:41
  SPEAKER_01: Some of that is genuine, and some of it is branding.
• 9:44
  SPEAKER_00: Right.
• 9:44
  SPEAKER_00: AI is the polite excuse.
• 9:47
  SPEAKER_00: Saying we are restructuring for efficiency sounds visionary, while saying our sales fell sounds like failure.
• 9:53
  SPEAKER_00: So I would bet some executives are dressing up ordinary cost cutting in futuristic language for investors.
• 9:58
  SPEAKER_01: But here is where I will challenge you a bit.
• 10:00
  SPEAKER_01: The beige book contacts are not talking to investors.
• 10:04
  SPEAKER_01: They are talking confidentially to their regional Fed.
• 10:07
  SPEAKER_01: So they have much less incentive to perform.
• 10:10
  SPEAKER_01: When a Cleveland banker privately says AI reduces staffing needs, I am inclined to take that at face value.
• 10:16
  SPEAKER_00: That is a good point, and I will concede it.
• 10:19
  SPEAKER_00: A confidential conversation with a Fed economist is a very different setting than an earnings call.
• 10:24
  SPEAKER_00: So the signal here is probably cleaner than the layoff trackers that pick up whatever companies choose to announce publicly.
• 10:30
  SPEAKER_01: There was another detail that made this feel very concrete for me.
• 10:34
  SPEAKER_01: The coverage mentioned a Dallas staffing company that used AI agents to conduct more than 180 hours of candidate interviews in just two days, with employees reviewing the results afterward.
• 10:44
  SPEAKER_00: That one is fascinating and slightly dizzying because it means AI is now on both sides of the hiring table.
• 10:51
  SPEAKER_00: Candidates are using AI to blast out hundreds of applications, and employers are using AI to interview them.
• 10:58
  SPEAKER_00: And somewhere in the middle, a human is supposed to make a judgment.
• 11:01
  SPEAKER_01: And look at what that does to recruiting jobs specifically.
• 11:04
  SPEAKER_01: 180 hours of interviewing is a full month of work for one recruiter, done in two days.
• 11:10
  SPEAKER_01: That is exactly the kind of productivity gain the beige book is describing, showing up inside the very function that hires people.
• 11:17
  SPEAKER_00: So let me bring this to the practical takeaway, because that is where I always want to land.
• 11:22
  SPEAKER_00: If you are a manager listening to this, what should you actually be doing differently on Monday morning, given that the Fed itself is flagging this trend?
• 11:30
  SPEAKER_01: First, be honest internally about whether you are choosing not to backfill because of AI and track it.
• 11:36
  SPEAKER_01: Because if you do not measure it, you cannot plan around it.
• 11:39
  SPEAKER_01: Second, think hard about your pipeline.
• 11:42
  SPEAKER_01: Because if you stop hiring juniors today, you will not have seniors in five years.
• 11:46
  SPEAKER_00: That second point is the one I would underline twice.
• 11:49
  SPEAKER_00: Every experienced engineer, accountant, or analyst I have worked with learned the craft by doing the boring entry-level work that AI now does.
• 11:57
  SPEAKER_00: So if that apprenticeship disappears, companies are quietly eating their own future talent bench.
• 12:02
  SPEAKER_01: And for workers, especially early career ones, the signal is to get fluent in supervising and checking AI output, not just using it.
• 12:11
  SPEAKER_01: Governor Cook mentioned that nearly half of small employers already use AI, and 71% of those report higher productivity.
• 12:18
  SPEAKER_01: So the expectation is spreading fast.
• 12:21
  SPEAKER_00: And for policymakers, I think the message is that the official statistics are going to lag this.
• 12:26
  SPEAKER_00: If the change shows up as jobs that never existed, the unemployment rate might stay calm for a long time while a whole generation of graduates quietly struggles to get started.
• 12:36
  SPEAKER_01: Which is why I would watch three things over the next few months.
• 12:39
  SPEAKER_01: The next payroll report, the next beige book in early December to see whether five districts become seven, and the big tech earnings calls later this month, where executives will tell us how much of their savings came from headcount.
• 12:52
  SPEAKER_00: Good list.
• 12:53
  SPEAKER_00: My own one-line summary is that AI is not firing people in large numbers yet.
• 12:58
  SPEAKER_00: It is just not hiring them.
• 12:59
  SPEAKER_00: And that quieter version of disruption is harder to see, harder to measure, and much harder for traditional policy tools to fix.
• 13:07
  SPEAKER_01: I agree.
• 13:08
  SPEAKER_01: And I think the significance of this beige book is that the Fed's own network of contacts is now saying out loud what many hiring managers have been doing quietly for months.
• 13:17
  SPEAKER_01: Once it is in the beige book, it is no longer a rumor, it is a data point.
• 13:21
  SPEAKER_00: That's all for today's episode of the DX Today podcast.
• 13:25
  SPEAKER_00: Thanks for listening, and we'll see you next time.