How will autonomous agents change company headcount and hiring?

Elias Torres · Mar 21, 2025 · hedged

We are entering a world where smaller is better — companies can potentially reach billions in revenue with far fewer people

Every prior generation assumed larger was better only because you couldn't reach billions in revenue without more people; that constraint is lifting

Scope: framed as the potential rather than a certainty

20:14 20VC: Selling Drift for $1.2BN is the Biggest Failure: What No One Tells You About Selling Your Company | Why Incumbents Are Slower & Worse Than Ever | Why the Most Valuable Companies in a World of AI Will Not Have More Than 100 People with Elias Torres

Elias Torres · Mar 21, 2025

The claim that AI won't reduce headcount because companies will just become more productive and do more things is false; companies will not keep growing their workforces

If we're building on the promise that AI does the work of humans, you can't simultaneously promise to hire a thousand more sales reps

24:10 20VC: Selling Drift for $1.2BN is the Biggest Failure: What No One Tells You About Selling Your Company | Why Incumbents Are Slower & Worse Than Ever | Why the Most Valuable Companies in a World of AI Will Not Have More Than 100 People with Elias Torres

Kieran Flanagan · Jul 11, 2025

Every go-to-market function will be smaller in the future and teams will be much, much smaller

Being honest about AI's effect, even while trying to frame the most optimistic view for people worried about their jobs

13:29 20Growth: The Death of Growth Teams? | How Hubspot Use AI to Triple Email Conversion | The Future of AI SEO | Why Prompt Engineering is the New Coding | What Every CMO Needs to Know About AI in 2025

Josh Browder · May 18, 2026

Companies of the future will be dramatically smaller in headcount

At DoNotPay customer support, work that once took three people now takes one, and even their ~10 contractors may be too many now that agents identify issues and pre-process refunds

68:29 20VC: Turning Peter Thiel's $100K into $10M Angel Portfolio | The One Man Accelerator at The Four Seasons | Why VCs Can Be Sharks and What Founders Need to Know | Why Stocks and Cash are BS and You Should Invest in Land with Josh Browder

Mike Mignano · Jul 6, 2026

Engineering organizations will evolve toward smaller teams of higher-caliber engineers

Lower-level tasks can increasingly be delegated to agents

25:33 20VC: Why Now is the Time for the Application Layer | Why OpenAI & Anthropic Won't Win the App Layer | Why Startups Should be TokenMaxxing | Why VCs Should Reduce Weighting on Price & Ownership in an Age of AI with Mike Mignano, USV

Fred Turner · Jul 18, 2026

Entire back-office departments in health insurance can be reduced to zero headcount with in-house agents, with dramatic cost and turnaround improvements

Their credentialing process — license checks, transcripts, malpractice databases, approval — went from 2-3 months and ~$50 per case with 5-6 staff to ~12 hours and ~20 cents using an in-house agent running on Claude

Scope: rubber-stamp-style, rule-following processes

41:38 20VC: $5BN in Revenue, 7 to 7,000 Employees in 9 Months, 206,000 Tests in a Single Day: The Craziest Story in Startups: Curative with Fred Turner

Fred Turner · Jul 18, 2026

AI is collapsing the headcount needed for nuclear engineering design, letting ~20 strong mechanical engineers do work that historically took ~100

Models have gotten really good at mechanical design, so humans focus on the critical pieces and oversee AI on routine parts like a bracket joining two pieces

Scope: humans still design and oversee the important/critical pieces

75:49 20VC: $5BN in Revenue, 7 to 7,000 Employees in 9 Months, 206,000 Tests in a Single Day: The Craziest Story in Startups: Curative with Fred Turner

Matt Swulinski · Aug 15, 2026

Marketing teams today and going forward should be far leaner — functions that used to need a team of three to five people can be run by one excellent person with a suite of agents

An agentic system lets one person operate at a scale (e.g. 100,000 influencers posting in a month) that previously required a team, so the 1% hit rate on candidates is acceptable

Scope: assumes the person can build the systems and knows where to plug into the data

59:42 20Growth: How to Build a $100M Growth Engine: Lessons from Wispr Flow and Superhuman | Why You Should Do Paid Ads Today and How To Do Them | How to Build the Best Referral Programs and How to Crush UGC with Matt Swulinski

Kyle Norton · May 30, 2025 · hedged

AI will not cause mass unemployment; it may instead lead to high employment

Jevons paradox — as compute got cheap usage exploded rather than collapsing in value; if each salesperson and IC becomes far more effective, demand for them rises, and AI will drive far more company creation needing those roles

Scope: admits he cannot read the future on this

72:49 20Sales: From $2M ARR to $40M ARR: The Playbook | How to Use AI To Supercharge Your Sales Team | Why Pipeline Reviews are BS | The Sales Call Script that Closes 99% of Prospects and How to Hire the Best Sales Reps with Kyle Norton, CRO @ Owner

Amit Bendov · Sep 12, 2025

Gong will have more engineers in ten years, not fewer, because business growth will outpace AI coding productivity gains.

Vibe coding delivers maybe 15% productivity improvement today, but the business is growing much faster, so seizing the opportunity requires more engineers.

Scope: current AI productivity gain estimated at ~15%; conditional on seizing the growth opportunity

60:16 20VC: Why AI SDRs are BS and Do Not Work | How to Use AI in Your Sales Team and Process to Win Today | What Skills Do All New Reps Need to Have in an AI First World with Amit Bendov, CEO @ Gong

Alexander Embiricos · Feb 21, 2026

Coding is one of the first domains where LLMs are genuinely strong, but calling coding 'automated' is misleading — historically each layer of automation expanded demand for the output rather than eliminating the work

Moving from assembly to higher-level languages didn't mean coding was automated; it let us write far more code and increased demand for engineers, just as the mechanization of human 'computers' and tabulation work exploded demand for that output

Scope: specific tasks do get automated; the terminology for the role changes over time

5:23 20VC: Codex vs Claude Code vs Cursor: Who Wins, Who Loses | Will All Coding Be Automated - Do We Need PMs | The Real Bottleneck to AGI | The Three Phases of Agents and What You Need to Know with Alex Embiricos, Head of Codex at OpenAI

Nikesh Arora · Jun 22, 2026

The belief that AI will mean fewer people working overall is a fallacy — companies will need more technical and more sales resources, not fewer

Every team wants more AI-savvy people to prompt frontier models, build harnesses and bring proprietary data to bear, plus more compute and storage; and if the product is good you need more salespeople because customers still don't know what you offer

Scope: applies to technical and sales roles, in contrast to his G&A headcount prediction

16:44 20VC: Nikesh Arora on the Frontier Model Problem: Breadth vs Depth | The Future of Token Costs | Memory Becoming the Moat | Where Value Accrues: Infra, Models, or Apps? | Why Enterprise AI is Not Ready & Systems of Record vs Systems of Intelligence

Arvind Jain · Jul 11, 2026

Engineering teams will not shrink: per-person productivity will rise sharply but so will demands, so making the same revenue will require producing a 10x better product.

Rising productivity is matched by rising competitive expectations.

33:34 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Arvind Jain · Jul 11, 2026

Composite roles are compatible with not shrinking headcount because companies will have to do roughly ten times the work to earn the same revenue, so smaller teams are simply forced to deliver more

The work required per dollar of customer revenue is rising sharply in the future

41:22 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Julien Bek · Aug 24, 2026

AI is increasing rather than reducing software engineering employment, via Jevons paradox — cheaper building means people build far more

The data shows more software engineers hired than ever; making building cheaper unlocks the enormous backlog of human creativity and ideas

64:04 20VC: Inside Sequoia's Investment Committee: Lessons from Don Valentine, Doug Leone and Alfred Lin | How the SpaceX and Citadel Deals Went Down | What Sequoia Specifically Looks for in Founders with Julien Bek

Harry Stebbings · Jan 12, 2026

Customer support labor is already clearing out, as shown by companies like Decagon and Harvey

64:54 20VC: a16z's $15BN Fundraise with Alex Rampell | The Best Companies Have Hostages Not Customers | The Best Founders Materialise Capital, Customers and Labour | Mid-Sized Funds with Die and The Future of Venture Capital

Ryan Petersen · Jun 20, 2026 · hedged

Flexport is hiring many salespeople and expanding into new markets, and will shift people from manual operations work toward customer-facing account management and sales

Operations work is being automated, and staff who can't make the transition will need to be rebalanced into customer-facing roles.

20:43 20VC: Why Remote Work is White Collar Fraud | Why Revenge and Patriotism are the Best Founder Traits | Two Questions Every Founder Needs to Ask | The Wild Story of Raising $1BN from Masa Son in an Hour Long Meeting with Ryan Peterson, Founder @ Flexport

Arvind Jain · Jul 11, 2026

Data and business analyst roles that are execution-only rather than business thinking will disappear, as business owners get answers directly, and recruiting sourcer roles will be absorbed into full-cycle recruiting

Dashboard-building and back-end configuration work is exactly what AI removes; business intelligence becomes self-serve

Scope: applies to analysts who are not business thinkers

41:43 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Fred Turner · Jul 18, 2026

Fears of AI job loss are warranted for a lot of back office roles.

Scope: specific to back office jobs

54:59 20VC: $5BN in Revenue, 7 to 7,000 Employees in 9 Months, 206,000 Tests in a Single Day: The Craziest Story in Startups: Curative with Fred Turner

Fred Turner · Jul 18, 2026

The two areas worth continuing to invest in people are technical skills and relationships; these won't be displaced soon.

Engineers deploying AI are dramatically more productive and still needed, and health insurance has irreducibly relationship-driven parts — members want a human somewhere in the loop, provider groups need relationship management, and brokers want finalist presentations, dinners and golf.

Scope: particularly true of health insurance; 'anytime soon' timeframe

55:17 20VC: $5BN in Revenue, 7 to 7,000 Employees in 9 Months, 206,000 Tests in a Single Day: The Craziest Story in Startups: Curative with Fred Turner

Fred Turner · Jul 18, 2026

Clinical care navigator headcount should scale linearly with membership even in an AI-heavy company, because members need a human point of contact

Agents let navigators manage a population at scale and handle follow-ups, so the same headcount per member does far more

Scope: specific to clinical/care-navigation workflows

59:51 20VC: $5BN in Revenue, 7 to 7,000 Employees in 9 Months, 206,000 Tests in a Single Day: The Craziest Story in Startups: Curative with Fred Turner

Bret Taylor · Oct 2, 2024

New job categories will emerge around agents — the 'agent engineer' who builds them and the 'AI architect' who does conversation design and shapes agent behavior — as meaningful as UI designer or web developer were for the web.

Just as the web took off and user experience design became a new category of job beyond boxed software, agents create their own craft roles.

52:33 20VC: Bret Taylor: The AI Bubble and What Happens Now | How the Cost of Chips and Models Will Change in AI | Will Companies Build Their Own Software | Why Pre-Training is for Morons | Leaderships Lessons from Mark Zuckerberg

Tobias Lütke · May 4, 2026 · hedged

A 'product builder' role will emerge as an ascension path from product, design or engineering, defined by coordinating intelligent actors — human and AI — to build products

Managers are increasingly best deployed prompting both people and AIs as part of the same team

Scope: it won't be named or described the way he describes it

53:01 20VC: Shopify's Tobi Lütke on How AI is a Scapegoat for Mass Layoffs & What Will Labour Markets Be in the Future | Why We Need More Scrutiny on Charitable Giving, Governments are Bad at What They Do and Trump Derangement Syndrome in Canada

Andrew Feldman · May 26, 2026

AI will spawn entirely new corporate roles around AI governance, just as networking created the CIO and security created the CSO, while other roles disappear entirely

Every major technical transformation of the past 25-30 years has produced jobs and organizations that didn't exist before and eliminated others — the VP of telco infrastructure role vanished completely along with PBX vendors

Scope: uncertain whether 'chief AI officer' is the right form

33:00 20VC: Cerebras CEO on the Future of Data Centres, Token Costs and Memory | We are Not in an Infra Bubble & Dario Got a Bad Deal with Elon for Compute | Should US Companies Sell to China & Why Most Layoffs are AI Washed with Andrew Feldman

Brendan Foody · Jun 1, 2026

Training agents will become a massive job category, and all knowledge work is converging on training agents rather than doing the work directly.

It is structurally more efficient to teach an agent a workflow once and amortize it over its useful life than to redundantly repeat the task — true for support reps, lawyers doing redlines, or anyone drafting emails and prepping meetings.

17:26 20VC: Mercor CEO on Why Application Layer Companies Have No Defensibility, The Model is the Product | Token Spend Will Exceed Headcount Spend in 5 Years | The True Cost of Hiring AI Researchers in the Valley Today with Brendan Foody

Harry Stebbings · Jun 20, 2026

Agents will in large part replace humans, creating a new corporate function of 'agent resources'

Unlike humans, agents genuinely are resources you can mine; and agents work a 007 schedule (midnight to midday, seven days a week) rather than 996

44:14 20VC: Why Remote Work is White Collar Fraud | Why Revenge and Patriotism are the Best Founder Traits | Two Questions Every Founder Needs to Ask | The Wild Story of Raising $1BN from Masa Son in an Hour Long Meeting with Ryan Peterson, Founder @ Flexport

Fred Turner · Jul 18, 2026

'Agent supervisor' will be a major job that doesn't really exist today

Agent workflows generate a high volume of exception and approval requests — even a 1% exception rate is 10x more requests when you do 10x the work — and managing those nuanced decisions without rubber-stamping is a bottleneck

Scope: agents supervising agents works to a degree; more expensive models may serve as supervisors

62:22 20VC: $5BN in Revenue, 7 to 7,000 Employees in 9 Months, 206,000 Tests in a Single Day: The Craziest Story in Startups: Curative with Fred Turner

Aaron Levie · May 22, 2024

Mature companies will reinvest AI productivity gains into the same functions rather than cut headcount, because there is more work to be done than they have ever had people to do

Empirically at Box, every productivity gain so far has been reinvested; freed customer support time gets redirected into proactive customer success

Scope: based on Box's own experience so far; some companies will instead pocket the efficiency dollars

17:11 20VC: Box's Aaron Levie on Predictions for the Next Wave of AI: Will Foundation Models Be Commoditised | How the Business Model of SaaS Changes Forever | Startups vs Incumbents: Who Wins | App vs Infrastructure Layer: Where is the Value?

Kieran Flanagan · Jul 11, 2025

Smart companies will redeploy the time AI gives back into higher-value human work rather than just banking better CAC and unit economics

AI's core output is time back, and how that time is redeployed is a company choice

14:14 20Growth: The Death of Growth Teams? | How Hubspot Use AI to Triple Email Conversion | The Future of AI SEO | Why Prompt Engineering is the New Coding | What Every CMO Needs to Know About AI in 2025

Alex Rampell · Jan 12, 2026 · hedged

Even where AI decimates a labor category like customer support or legal grunt work, smart companies will reallocate that labor to high-touch relationship work rather than simply shrinking headcount

Once the cost-center element is automated away, spending on making customers love you becomes a revenue driver — a United could give best travelers a dedicated human who sells more first class tickets

Scope: the tedious-work headcount does fall off a cliff; applies to smart companies, not all

65:03 20VC: a16z's $15BN Fundraise with Alex Rampell | The Best Companies Have Hostages Not Customers | The Best Founders Materialise Capital, Customers and Labour | Mid-Sized Funds with Die and The Future of Venture Capital

Arvind Jain · Jul 11, 2026

Companies that shrink headcount because AI lets them do the same work with fewer people will be beaten by competitors who keep their people and use AI to produce a 10x better product or 10x more output.

Both firms have equal access to AI, so the one that redeploys headcount toward more ambitious output rather than cutting will win.

27:43 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Arvind Jain · Jul 11, 2026

The world's greatest companies will not be 100-person companies; people remain an asset that must be deployed onto the right projects.

The post-COVID cuts showed teams get bloated and slow each other down, but that dynamic predates AI; and the model companies themselves are hiring aggressively.

Scope: accepts that large teams do slow each other down

29:04 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Tobias Lütke · May 4, 2026

AI will be blamed for absolutely everything because it is the perfect girardian scapegoat — it can't fight back.

It cannot defend itself, and there is no constituency on AI's side; the tech industry has been gaslighting people into AI fear and sci-fi did so for sixty years before that.

18:53 20VC: Shopify's Tobi Lütke on How AI is a Scapegoat for Mass Layoffs & What Will Labour Markets Be in the Future | Why We Need More Scrutiny on Charitable Giving, Governments are Bad at What They Do and Trump Derangement Syndrome in Canada

Josh Browder · May 18, 2026 · hedged

Big tech layoffs are AI-driven at Meta but mostly COVID-era overhiring at companies like Block

Meta's heavy data center spend has made its engineering more productive so it needs fewer engineers; Block's case looks like hiring correction

Scope: 'maybe it's controversial'; depends on the company

68:09 20VC: Turning Peter Thiel's $100K into $10M Angel Portfolio | The One Man Accelerator at The Four Seasons | Why VCs Can Be Sharks and What Founders Need to Know | Why Stocks and Cash are BS and You Should Invest in Land with Josh Browder

Andrew Feldman · May 26, 2026

90-95% of layoffs to date have been AI-washed — they are actually the result of boneheaded COVID-era overhiring and the harvesting of productivity gains accumulated over years, not AI

The ability to gather and synthesize information across an organization has been changing middle management for years; AI is only just now starting to have meaningful enterprise impact, but it is an easy umbrella to put terminations under

Scope: in my view; AI is only just beginning to have real enterprise impact

30:20 20VC: Cerebras CEO on the Future of Data Centres, Token Costs and Memory | We are Not in an Infra Bubble & Dario Got a Bad Deal with Elon for Compute | Should US Companies Sell to China & Why Most Layoffs are AI Washed with Andrew Feldman

Arvind Jain · Jul 11, 2026

The argument that leaner teams move faster is not an AI argument — it has always been true.

28:48 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Alexander Embiricos · Feb 21, 2026

There will be many more builders in five years, not fewer, but the talent stack will compress — engineers will be far more full-stack and role boundaries like front-end vs back-end will fade

He observes this compression already on the Codex team, where the front-end/back-end split is much less pronounced than a few years ago

Scope: terms like 'engineer' may come to mean something different

6:36 20VC: Codex vs Claude Code vs Cursor: Who Wins, Who Loses | Will All Coding Be Automated - Do We Need PMs | The Real Bottleneck to AGI | The Three Phases of Agents and What You Need to Know with Alex Embiricos, Head of Codex at OpenAI

Harry Stebbings · May 16, 2026

The single best advice for a graduate today is to be full-stack — e.g. a marketer who writes the copy and edits the video themselves

56:45 20VC: Lessons from Jensen Huang on "Founder Mode" | How to Know if OpenAI or Anthropic Will Kill your Company | How USV Liking Music Made Them $1BN on an Investment | The Five Year Desert to Product Market Fit & a $5.3BN Valuation with Shiv Rao @ Abridge

Harry Stebbings · Jun 13, 2026

Every function, not just engineering, now demands full-stack ownership from start to finish rather than handoffs between specialists

In marketing, for example, you should create the copy, ready it, post it at the right time and amplify it, rather than handing off to designers and a social team

32:22 20VC: Who Wins the Model War: OpenAI, Anthropic or Open-Source | Token Maxing, AI Hangovers & The Coming ROI Reckoning | Labour Displacement Fears are BS & Overblown | From Physicist to Sequoia Founder with Matan Grinberg, Founder @ Factory

Arvind Jain · Jul 11, 2026

Composite generalist roles will become common in three to five years — one person doing engineering, product and design, or selling, demoing and solutioning — reversing the trend toward specialization

AI lets individuals cover what used to require separate specialist functions

Scope: engineering and go-to-market examples

40:29 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Daniel Khachab · Oct 28, 2024

Great companies will get smaller in headcount rather than bigger, and a ten-person (or even one-person) unicorn will happen

You have to automate as much as possible and AI now makes that largely feasible

Scope: may be ~10 people rather than literally one; says it is 'probably' possible today with the right 10 people

63:34 20VC: Why SaaS is Dead | Why AI First Companies Will Win | We are in the Middle of a Cold War for AI Talent | Why Europe is F******* and We Need to Stop Whining with Daniel Khachab, Co-Founder @ Choco

Edwin Chen · Jul 21, 2025

A single-person billion-dollar company will exist one day

10x and even 100x engineers already exist, single-person startups are already doing $10M in revenue, and AI efficiency gains could multiply that by 100x

Scope: framed as eventually, not imminently

10:38 20VC: Scaling to $1BN+ in Revenue with No Funding: Surge AI | The Most Insane Scaling Story in Tech |

Alex Schultz · Sep 5, 2025 · speculative

If AI intelligence advances as the leaders of the field are suggesting, things get very weird and you will see entire companies run by one person

The scale of intelligence advances being promised implies far more collapsing of roles than merging product, growth and engineering into one person

Scope: conditional on AI advances materializing as field leaders suggest

12:38 20Growth: Meta CMO Alex Schultz on How All Founders Have to Change Their Marketing Playbook in a World of AI | Is AI Plateauing, What it Means if China Wins the AI Race and Why Zuck is a Generational Leader

Kieran Flanagan · Jul 11, 2025

AI gives companies a choice between two uses: doing the same work with fewer people to save capital, or redeploying the same people to do more and grow faster

Low-complexity support and BDR work can be absorbed by AI, so the freed capacity can be banked as cost savings or reinvested into white-glove onboarding and more meetings

11:59 20Growth: The Death of Growth Teams? | How Hubspot Use AI to Triple Email Conversion | The Future of AI SEO | Why Prompt Engineering is the New Coding | What Every CMO Needs to Know About AI in 2025

Amjad Masad · Apr 25, 2026

AI does not uniformly shrink companies — some entrepreneurs use it to stay lean while others reinvest the efficiency gains into hiring more people

He sees both patterns among customers: Jason runs leaner and does more, while a board-game seller hired eight people and Firecrown Media hired more vibe coders after saving on SaaS and selling more

Scope: depends on the kind of company, the entrepreneur and their level of ambition

33:06 20Product: Replit CEO on Why Coding Models Are Plateauing | Why the SaaS Apocalypse is Justified: Will Incumbents Be Replaced? | Why IDEs Are Dead and Do PMs Survive the Next 3-5 Years with Amjad Masad

Matan Grinberg · Jun 13, 2026 · hedged

Whether AI leads to smaller teams or same-sized teams with bigger ambitions is genuinely unpredictable; each company will choose between solving the same problem, a bigger problem, or more problems

The dynamics are hard to predict; it depends on how each business chooses to deploy its new leverage

Scope: declines to predict a direction

6:24 20VC: Who Wins the Model War: OpenAI, Anthropic or Open-Source | Token Maxing, AI Hangovers & The Coming ROI Reckoning | Labour Displacement Fears are BS & Overblown | From Physicist to Sequoia Founder with Matan Grinberg, Founder @ Factory

Adam Foroughi · Apr 27, 2026

It is right to cut 40-50% of staff even while growing near triple digits if roles are going to be automated or a department isn't adopting AI fast enough; you should rebuild the organization as if you were building it today with the technology now available

Roles heading for automation and departments slow to adopt AI shouldn't be carried forward

17:16 20VC: Applovin: $160BN Market Cap, $5.48BN Revenue, $10M EBITDA Per Head | Why the Best Do Not Need Mentorship | Why Founders Should Not Angel Invest | Why Kindness in Business Will Slow You Down with Adam Foroughi

Adam Foroughi · Apr 27, 2026

Cutting roles before the automating technology has fully arrived is better than keeping people in roles heading for a dead end, even if the bet on the technology is not yet proven

Dead-end roles hit morale and make the organization optimize for unhappy people; A players want free rein and don't want to be distracted by unhappiness around them; severance frees those people to go do something else

Scope: Requires good severance; Acknowledged as a bet on technology getting good enough

19:54 20VC: Applovin: $160BN Market Cap, $5.48BN Revenue, $10M EBITDA Per Head | Why the Best Do Not Need Mentorship | Why Founders Should Not Angel Invest | Why Kindness in Business Will Slow You Down with Adam Foroughi

Adam Foroughi · Apr 27, 2026 · hedged

There will be a lot more tech layoffs over the next couple of years as companies realize that not laying people off blocks them from becoming AI-native

Every technology company could shed 75-80% of its talent and become more efficient, and keeping those people is itself the blockade to reaching an AI-native state

Scope: "my guess is"

63:55 20VC: Applovin: $160BN Market Cap, $5.48BN Revenue, $10M EBITDA Per Head | Why the Best Do Not Need Mentorship | Why Founders Should Not Angel Invest | Why Kindness in Business Will Slow You Down with Adam Foroughi

Jonathan Ross · Feb 17, 2025

In an inference-dominated world, compute becomes substitutable for headcount — a startup hires 10 people instead of 100 and buys compute equivalent to the other 90

An LPU or GPU is like an employee that can't quit and never needs retraining, so once a model hits a capability level it stays there, letting CapEx or OpEx replace labor

17:13 20VC: NVIDIA vs Groq: The Future of Training vs Inference | Meta, Google, and Microsoft's Data Center Investments: Who Wins | Data, Compute, Models: The Core Bottlenecks in AI & Where Value Will Distribute with Jonathan Ross, Founder @ Groq

Elias Torres · Mar 21, 2025 · hedged

Smaller teams may not reduce total capital needs, because companies are trading people for compute and compute still has to be funded

Headcount spend is being substituted with compute spend rather than eliminated

Scope: "maybe not" — offered tentatively

23:16 20VC: Selling Drift for $1.2BN is the Biggest Failure: What No One Tells You About Selling Your Company | Why Incumbents Are Slower & Worse Than Ever | Why the Most Valuable Companies in a World of AI Will Not Have More Than 100 People with Elias Torres

Harry Stebbings · Jul 11, 2026

Cutting headcount to fund the best frontier models and tooling produces better products faster, because top engineers want to join companies with the best technology.

Reduced headcount frees budget for frontier tools, which attracts the best engineers; more people does not make better products.

28:26 20VC: Why OpenAI and Anthropic Won't Win the App Layer | Why Teams Will Get Bigger Not Smaller in a World of AI | Why AI Removes Incumbents Advantage of Bundling | China vs America: Who Wins the AI War with Arvind Jain, Co-Founder @ Glean

Tom Hulme · Apr 10, 2025

The idea of a one-person billion-dollar business is a myth

Just as foundation models are being distilled, business models and businesses will be distilled and copied ridiculously quickly, so no single person can hold a moat wide enough for a billion dollars of revenue

Scope: accepts companies are growing faster and more efficiently than ever, e.g. dot new at $40M run rate in three months

68:23 20VC: How to Fix the UK Tech Ecosystem | Why We Need to Flood the UK with Venture Capital | What the UK Can Learn From Sequoia, Stripe and Norway | Why Now is the Time to be Bullish on China & Lessons from Jensen Huang with Tom Hulme & Stan Boland

Varun Mohan · Jun 2, 2025

Single-person billion-dollar companies will not happen

We live in a competitive capitalist market — unless you're a thousand times smarter than everyone, two equally smart people will copy your idea, compressing margins, brand pervasiveness and customer acquisition costs; a solo billion-dollar product would imply nobody else cares about the opportunity

Scope: absent continued investment

52:22 20VC: Windsurf Founder on Will Model Companies Own the App Layer | Why Moats Do Not Exist in a World of AI | Why the Notion of Single Person $BN Companies is BS | Lovable vs Bolt & Cursor vs Windsurf: How Does it All End with Varun Mohan

Varun Mohan · Jun 2, 2025

If AI became good enough that a solo founder needed no employees, competitive pressure would increase rather than decrease, because everyone else would have the same AI

A company is the sum of discounted cash flows over time and will eventually need people; if AI does everything, that capability is universally available and the founder's own contribution is unclear

53:13 20VC: Windsurf Founder on Will Model Companies Own the App Layer | Why Moats Do Not Exist in a World of AI | Why the Notion of Single Person $BN Companies is BS | Lovable vs Bolt & Cursor vs Windsurf: How Does it All End with Varun Mohan

Amit Bendov · Sep 12, 2025

Go-to-market orgs will consolidate into fewer, blended roles — a 'one person band' where a seller also does customer success — because agents give individuals the tools to cover multiple jobs

Companies that aren't hypergrowth don't need structures optimized for hypergrowth, and agents let one person do several jobs

32:55 20VC: Why AI SDRs are BS and Do Not Work | How to Use AI in Your Sales Team and Process to Win Today | What Skills Do All New Reps Need to Have in an AI First World with Amit Bendov, CEO @ Gong

Anish Acharya · Feb 9, 2026

The big productivity unlock is bundling support, sales, collections and operations into one function aimed at a broad goal like CAC improvement — not taking cost out of customer support in isolation

Enterprises were organized around two human archetypes — the empathetic listener for support and the charismatic yapper for sales — but models can be either archetype at any time, so the organizational split no longer needs to exist

Scope: describes it as what the most sophisticated companies are starting to do

32:41 20VC: Is SaaS Dead in a World of AI | Do Margins Matter Anymore | Is Triple, Triple, Double, Double Dead Today? | Who Wins the Dev Market: Cursor or Claude Code | Why We Are Not in an AI Bubble with Anish Acharya @ a16z

Harry Stebbings · Jul 11, 2025

Handing rote content production like social clips to AI frees media teams to do far more creative storytelling and show promotion

The clips are rote and near-identical yet take days, crowding out time for creative storytelling

Scope: based on his own media team

13:48 20Growth: The Death of Growth Teams? | How Hubspot Use AI to Triple Email Conversion | The Future of AI SEO | Why Prompt Engineering is the New Coding | What Every CMO Needs to Know About AI in 2025

Andrew Macdonald · Aug 17, 2026 · hedged

Uber could do everything it does today with fewer people in five years, but because it will be doing a lot of new things, total headcount may still need to grow

Producing-type functions (support, sales, content, analytics) lend themselves to AI augmentation and eventually partial replacement; but predictions of AI-driven headcount decline have been repeatedly proven wrong and new fields emerge as in past industrial revolutions

Scope: 'fewer people' holds only if the business is held static; he declines to state the reduction emphatically

42:02 20VC: Uber President on The Untold Uber Stories: Travis, China and Self-Driving | Why Autonomy Is Existential | How to Beat DoorDash to #1 in Food with Andrew MacDonald

Kieran Flanagan · Jul 11, 2025

The future of marketing and knowledge work is a small number of 'super ICs' — AI-enabled generalists who cover what used to require multiple narrow specialists

AI assistance lets one generalist do much more and experiment and get feedback far faster, whereas teams that don't use AI simply learn more slowly

51:33 20Growth: The Death of Growth Teams? | How Hubspot Use AI to Triple Email Conversion | The Future of AI SEO | Why Prompt Engineering is the New Coding | What Every CMO Needs to Know About AI in 2025

Shiv Rao · May 16, 2026 · hedged

Companies can now be built far flatter — fewer managers and more super-ICs each taking on much more

AI tools let everyone move in lockstep and stay coordinated in a way that was impossible before

Scope: describes himself as in an idealistic mode; framed as near-future

56:17 20VC: Lessons from Jensen Huang on "Founder Mode" | How to Know if OpenAI or Anthropic Will Kill your Company | How USV Liking Music Made Them $1BN on an Investment | The Five Year Desert to Product Market Fit & a $5.3BN Valuation with Shiv Rao @ Abridge

Jason Wilk · Apr 21, 2025

Dave's swing to profitability came from AI-driven leverage in underwriting and support plus contract cleanup, not from headcount reduction

Headcount is the same 300 people as at IPO with no RIF; AI cut support costs and, more importantly, unlocked leverage in underwriting

27:34 20VC: Do Rich Founders Make Better Founders | The Best Performing Fund Would Only Back YC Founders on Their Second Time | Why SPACs Will Come Back | Why Short Sellers Should Be Banned | Is Trump Better for Business than Biden with Jason Wilk @ Dave

Byron Deeter · Aug 25, 2025

Growing revenue while shrinking headcount is not ruthless leadership — executives are giving employees tools to do the strategic, creative work while the business grows without the workforce growing

The stated posture of these executives is that tech works for you rather than you being a slave to tech, supercharging what everyone is doing

21:08 20VC: Do Margins Matter in AI? | Is Defensibility Gone For Good? | Is Vertical SaaS Dead in a World of AI | What SaaS Rules Are BS and No Longer Apply in a World of AI | The Future of Venture: Why Chanel vs Walmart is BS with Byron Deeter

Jake Saper · Mar 10, 2025 · hedged

Enterprises deploying AI today are not firing people; they are declining to hire new people and growing output with existing headcount

Customer interviews at a healthcare voice AI company showed zero headcount reduction but 3x business growth on the same headcount; buyers also emotionally resist firing staff

Scope: based on what he has seen so far; still early days

57:02 20VC: Lessons from Investing $2BN and Returning $8BN in Cash | Why Most Venture Partnerships are Broken | We Sold Salesforce Early and Lost Out on Billions | Are The Best Deals Always Expensive and Competitive with Jake Saper @ Emergence Capital

Jerry Murdock · Feb 28, 2026

The first labor market effect of autonomous agents is not firing existing workers but slowing or stopping hiring of the next junior developer, executive assistant or marketing person

Anyone who inputs data into a computer will ultimately be better served by an autonomous agent, and the direct threat lands on the next person in line rather than the incumbent

25:42 20VC: Why Cursor is Dead | An AI Tsunami is Coming & You Need to Prepare | Systems of Record Become Valueless Databases with Agents | Is This The End of Tech Private Equity with Jerry Murdock, Co-Founder of Insight Partners

Alex Rampell · Jan 12, 2026

The belief that more input produces better output is wrong; a much smaller team of exceptional people beats a large organization

Smaller teams require less communication, are forced into creative and technological solutions, and there is addition by subtraction — but only if the people are great

Scope: requires the small team to be top-tier talent, not just fewer people

34:02 20VC: a16z's $15BN Fundraise with Alex Rampell | The Best Companies Have Hostages Not Customers | The Best Founders Materialise Capital, Customers and Labour | Mid-Sized Funds with Die and The Future of Venture Capital

Jerry Murdock · Feb 28, 2026

Whether shrinking headcount is good depends entirely on culture — a company built on A players will end up with far more agents than people

Steve Jobs' framing at Apple was about continuing to grow with A players; the question is what the remaining people are doing and whether their quality of life improves

Scope: declines to judge Klarna specifically

30:15 20VC: Why Cursor is Dead | An AI Tsunami is Coming & You Need to Prepare | Systems of Record Become Valueless Databases with Agents | Is This The End of Tech Private Equity with Jerry Murdock, Co-Founder of Insight Partners

Aaron Levie · Apr 20, 2026

Because you can't blame the model provider when an agent causes harm, organizations must retain internal accountability and ownership chains, so AI does not dissolve existing hierarchies, contract law or regulatory regimes

Customers won't accept 'the system we set up screwed up your data' as an excuse, and any liability requires someone accountable who rolls up to someone with more accountability

32:33 Accountability and liability requirements preserve hierarchy

Harry Stebbings · Nov 24, 2025 · hedged

A public company is better off reporting a larger headcount because a less sophisticated investor base reads headcount as security rather than understanding AI efficiency in dev teams

Investors who don't think about AI efficiency within dev teams find bigger teams reassuring

54:14 Public investors read headcount as security so companies keep it high

Fred Turner · Jul 18, 2026 · hedged

Current layoffs are a mix of correcting 2021-2022 overhiring, companies chasing valuation boosts from cuts, and genuine workflow change from AI.

Markets reward layoff announcements, which incentivizes possibly bad behavior, but some workflows really are changing.

59:21 Layoffs mix overhiring correction market signaling and real ai

Fred Turner · Jul 18, 2026 · hedged

Curative will shrink quite a bit from its current ~650 headcount in the short term before growing larger again over five years.

They are not done automating the backend workflows.

59:40 Headcount dips through automation then grows again

Mitchell Green · Mar 28, 2025 · hedged

Lead Edge's software portfolio companies will not be wholly disrupted away by AI, and one-person companies are not arriving as fast as the market assumes

69:42 Believes full ai disruption and one person companies are slower than market assumes

Hemant Taneja · Sep 22, 2025 · speculative

There is a nontrivial probability that large organisations restructure over the next ten to fifteen years to be mostly AI agents with a small human core

A large enterprise already asked its consultancy to draw up a plan to go from 50,000 employees to 100,000 of which only 10,000 are human — provocative, but indicative of how leaders are thinking

18:05 Large organizations restructure to mostly ai agents with a small human core within 10 15 years

Christian Lanng · Sep 27, 2023

Demographics force a change: companies could not replace their shared-services workers one-for-one even if they wanted to, because young people won't take jobs typing formulas into Workday

The younger generation refuses that kind of work, so without changing the work experience we run out of people to do it

50:50 Demographic labor refusal forces automation

Harry Stebbings · May 23, 2026

Benioff's stance is that Salesforce no longer needs to hire developers but does want to hire salespeople

48:41 Engineering headcount shrinks while sales headcount grows

Eran Zinman · Mar 2, 2026 · hedged

Headcount reduction should be done gradually and responsibly through the transition rather than by slamming on the brakes, even though the long-run direction is much lower headcount

There's a transition period between where software is today and where it's going

19:35 Reduce headcount gradually through the transition not abruptly

Elias Torres · Mar 21, 2025 · hedged

AI will force smaller teams because human headcount, not money, is the binding constraint on scaling support functions

At Drift he had 200 people in CS; public-company friends spending $30M on CS say the problem isn't the money but being unable to hire and train fast enough, plus the bloat of management layers, benefits, recruiting and onboarding — compute running 24/7 avoids all of it

20:52 Headcount not money is the binding constraint on scaling support functions

Aaron Levie · May 22, 2024

Org structures will stay essentially the same under AI, with a layer of AI labor dropped into parts of the org chart rather than a redesign of the chart itself

The org chart codified ~70-80 years ago has survived the internet and remote work; AI swaps in for or augments people where more support is needed

15:12 Org chart remains unchanged ai labor drops into existing structure

Harry Stebbings · Jun 27, 2026

There is a paradox in the current market: the narrative says engineers are being replaced by tokens, yet every founder names hiring as their biggest problem

24:05 Hiring demand contradicts the displacement narrative

George Sivulka · Jan 22, 2025

AI will augment rather than replace people — it will increase the AUM of firms that adopt it and drive more employment, while shifting lower-cognition back, middle and some junior front-office tasks toward managing AI juniors

Excel didn't take jobs away, it changed what people had to learn; the same pattern will hold for AI, just as bookkeepers tabulating on paper disappeared without employment falling

32:48 Ai augments rather than replaces increasing aum and employment while shifting task mix

Max Junestrand · Jan 26, 2026

Law firms will be bigger, not smaller, because they consolidate — but each transaction will be run by fewer people, with more transactions overall and each person running more of them

Technology raises per-person throughput, as with the shift from physical data rooms staffed by people opening boxes to virtual data rooms

50:31 Consolidation grows firms while each deal uses fewer people

Harry Stebbings · Jan 26, 2026

AI will cause more labor displacement than people expect, and it will happen within the next 12 to 24 months

52:28 Displacement arrives faster and larger than consensus

Andrew Dudum · Apr 4, 2026 · hedged

Hims & Hers will likely have not many more than 2,000 to 3,000 employees by 2030

AI drives efficiency in product, engineering, accounting, finance and marketing, but the large variable labor force running a million square feet of pharmacy fulfillment with required doctor and pharmacy oversight can't be leveraged as much

10:53 Physical operations and regulated oversight cap headcount reduction

Avishai Abrahami · Jul 13, 2026

AI will not take Wix's customer support headcount to zero or anywhere near a handful of people

Wix serves 192 countries and that scale of operation cannot be covered by ten people

31:28 Operational scale and breadth keep support headcount large

Matan Grinberg · Jun 13, 2026

A GM-style role for former engineers — owning a business outcome end to end including marketing copy, product metrics and sales enablement — will become a common function that barely exists today

31:47 Engineer gm owning a business outcome becomes a standard role

Julien Bek · Aug 24, 2026

Record hiring of system integrators and forward deployed engineers confirms that human judgment and hands are still required, and going from copilot to autopilot does not remove the human

ServiceNow and others have never hired more system integrators; in companies like Auctor one person does the job ten did, but with a human still in the loop

62:10 Rising integrator hiring shows humans stay in the loop

Harry Stebbings · Sep 5, 2025

The idea of one-person companies is ridiculous because enterprise buyers require customer success, security, and in-person relationship-building that a single founder cannot provide

A large enterprise buyer wants to meet you to judge whether you're a good person, and a solo founder can't simultaneously handle CS, security and global travel to customers

12:55 Enterprise requirements for cs security and in person trust block one person companies

Alex Schultz · Sep 5, 2025

People won't be replaced by AI; they'll be replaced by someone doing their job using AI

The next generation of practitioners will find ways to blend human craft with AI tools

36:42 People are replaced by ai wielding competitors not by ai itself