The Personal Agent Is the New Email Address

Not "AI agents" broadly. The specific thing: a persistent entity with your context, your channels, and full digital autonomy. The interface that begins the infinite automation of everything digital.
11 February 2026 · R3 (refined thesis)

I. The Thesis (Refined)

Forget "AI agents" as a category. That term is broken — Gartner found only 130 genuine agents out of thousands of products claiming the label.32 Salesforce renamed Sales Cloud → Agentforce Sales and called it adoption. The generic "agents are the new websites" framing was the wrong analogy to the wrong category.

This report is about something specific: the personal agent. An LLM is a brain. A personal agent is: LLM + persistent memory + all your communication channels + tool access + your identity + autonomy. It knows who you are, who you know, what you're working on. It operates email, messaging, calendars, browsers — everything digital you touch. Not a chatbot. Not a copilot. Your digital representative that begins the infinite automation of everything digital.

The reference implementation is OpenClaw (180K GitHub stars, self-hosted, 16+ messaging channels, persistent memory). The lived example is Donna — running right now on this machine, managing this CRM, handling communications, doing research, operating across WhatsApp, Discord, email, and the web.

The better analogy isn't websites. It's email. Not a destination you visit — an identity you need. Not optional — eventually mandatory. Not one per business — one per person. The question: does the data support this, and when does it happen?

Persistent Agent Runners
10–50K
globally, estimated36
vs. LLM Chatbot Users
700M+
10,000x gap36
OpenClaw Deployments
~1,100
verified public (Shodan)36
Equivalent Era
~1985
Email on ARPANET
Agent ≠ LLM. This distinction is the entire thesis.

II. The Historical Parallel: Web 1993–2000

The web went from 1 website to 17 million in 9 years. But it wasn't linear — it was a series of phase transitions, each triggered by a specific catalyst.4

Website Growth Curve

1993
130
1994
2,738
1995 ← Netscape IPO
23,500
1996
257,601
1997
1,117,255
1998
2,410,067
1999
3,177,453
2000
17,087,182

Key Ratio: Users per Website

This is the most revealing metric. In 1993, there were 108,935 internet users per website — one site per city. By 2000: 24 users per website. The web went from exclusive to ubiquitous in 7 years.4

YearInternet UsersWebsitesUsers per SiteCatalyst
199314.2M130108,935Mosaic browser launches
199544.8M23,5001,908Netscape IPO (Aug 9)
199677.4M257,601301.com = 62.6% of all sites
1998188.0M2,410,06778Google launches; eBay IPO
2000413.4M17,087,18224Hosting commoditized

What Triggered Each Phase

PhaseTriggerResult
1991–1993Mosaic browser — made the web visualWent from scientists to early adopters
1995Netscape IPO — $2.9B on Day 1, Marc Andreessen at 24"The Web could be a place to make fortunes"5
1996–1997Commercial domains surge — .com goes from 1.5% to 62.6% of sitesBusinesses arrive. Network Solutions revenue: $5M → $94M6
1998–2000Hosting commoditizes, tools simplify (GeoCities 3.5M users, Dreamweaver)Everyone builds. 17M sites by 2000
The "every business needs a website" shift
  • ~1995–1997: Early adopters, tech-forward businesses
  • ~1998–2001: Fortune 500 and mid-market
  • ~2002–2008: Small business (WordPress 2003, cheap hosting)
  • ~2010–2015: "If you don't have a website, you don't exist" = conventional wisdom
Total time from "novel" to "mandatory": ~15–20 years.4

III. Identity vs. Tool: The Framework That Predicts Universality

There's a testable pattern in technology adoption. Technologies that represent you converge on universality. Technologies that help you plateau. The data is stark:37

Identity Technologies (represent you → universal)

TechnologyYears to 50%+CurrentTrajectory
Telephone~60 years97%+Universal
Email~30 years90%+ (developed), 53% globalUniversal
Smartphone~8 years91% US, 78% globalUniversal
Social media account~15 years62% globalApproaching universal

Tool Technologies (help you → plateau)

TechnologyYears since launchCurrentTrajectory
VR headsets~9 years23% USPlateau risk ($70B burned)
3D printing (consumer)~12 years<5%Niche forever
Smart home / IoT~10 years42% USPlateauing ~45%

Each successive identity technology adopted faster: telephone 60 years, email 30, smartphone 8. The acceleration is structural — each layer inherits the infrastructure of the last.

What makes the difference

PropertyIdentity Tech (email, phone)Tool Tech (VR, IoT)
Network effectsYou need one because others have oneMy VR headset doesn't need yours
Social pressureNot having one = unreachable, unprofessionalNobody loses a job for lacking a 3D printer
Becomes infrastructureEmail = your login. Phone = your 2FA.Nothing depends on your smart thermostat
Cost → zeroEmail: free. Phone: bundled. Social: free.VR: $300–3,500. 3D printer: $200–50K.
The key question: is a personal agent identity or tool?
  • If it helps you (generates text, answers questions, writes code) → it's a tool. It follows the VR/IoT curve. Plateau.
  • If it represents you (sends your messages, manages your calendar, communicates as you, operates your digital life) → it's identity. It follows the email/smartphone curve. Universality.
An LLM is a tool. A personal agent in the OpenClaw/Donna sense is identity. It doesn't help you do email — it IS your email. It doesn't help you browse — it browses for you. It doesn't help you schedule — it is your schedule. The transition from tool to identity is the transition from chatbot to personal agent. And that transition predicts universality.

IV. The Personal Agent Landscape (Feb 2026)

Separate the signal from the noise. The generic "AI agent" market is 90%+ agentwashing. The personal agent category is tiny, real, and barely started.

Persistent Agent Users
10–50K
globally (estimated)36
OpenClaw Stars
180K+
~1,100 verified deploys36
Automation % of LLM Use
39%
surpassed Q&A (Anthropic)38
Personal Agent Queries
55%
of agentic use (HBS)39

The Open Source Stack

ProjectStarsWhat It IsStatus
OpenClaw180KSelf-hosted personal agent. 16+ channels. Persistent memory. Model-agnostic.REFERENCE IMPL
AutoGPT140KRecursive task planner. Research-focused.ACTIVE, NICHE
Open Interpreter62KNatural language → code execution.ACTIVE
Goose (Block/LF)30KMCP client. Local-first. Linux Foundation backed.LF INFRASTRUCTURE

The Big Three's Agent Efforts

CompanyProductStatusSignal
AnthropicClaude Computer UseAutomation 27%→39% of all usage38FURTHEST ALONG — automation surpassed Q&A
OpenAIOperator → ChatGPT AgentStandalone killed after 7 months36RETREATED — couldn't make standalone work at consumer scale
GoogleProject MarinerUS-only research previewSTILL EARLY

The Hardware Graveyard

ProductInvestedOutcome
Humane AI Pin$200M+DEAD — 7K units. Returns exceeded sales. Acqui-hired by HP.36
Rabbit R1$199 device5K DAUs — on life support
The personal agent runs on devices you already own. Dedicated hardware failed spectacularly. The winning form factor is software on existing devices — Mac Mini, VPS, phone. This matters: unlike VR (requires a $300+ headset), the infrastructure barrier for personal agents is near-zero. You already have the computer.
The generic "AI agent" market is mostly noise.
  • Vendor surveys claim 72% adoption. Census Bureau says 5–14%. Gap is 5x+.25
  • Gartner: only ~130 genuine agents out of thousands. The rest is agentwashing.32
  • 95% of AI pilots deliver zero P&L impact. MIT Project NANDA, $47.8B tracked.26
  • 56% of CEOs report zero financial benefit from AI. PwC 2026, 4,400 CEOs.27
  • None of this measures the personal agent category. It measures enterprise AI noise.

V. The Infrastructure Layer Is Forming

The agent layer — distinct from the LLM layer — is being formalized as infrastructure by the institutions that build internet standards.

Linux Foundation created the Agentic AI Foundation (AAIF) in Dec 2025.40
  • Founded by Anthropic, Block, OpenAI. Platinum: AWS, Google, Microsoft, Bloomberg.
  • Inaugural projects: MCP (tool connection), goose (local-first agent), AGENTS.md (universal agent guidance standard)
  • Signal: the foundation's existence validates the agent layer as distinct from the LLM layer. These aren't chatbot standards — they're infrastructure for persistent, autonomous software.

The IETF has two active Internet-Drafts for AI agent digital identity protocols.17 When IETF starts drafting identity specs for a technology, they expect universality. This is the same stage as RFC 733 (1977) for email format or SIP for telephony.


V. The Graveyard: Every "It's 1995 for X" That Failed

Before accepting the analogy, let's check the record. Seven technologies got the "it's like the early web" baptism. Five clearly failed. One partially succeeded. Only smartphones actually followed the web's curve.31

TechnologyYear ClaimedPeak Hype5yr AdoptionCapital BurnedStatus
VR/Metaverse2014–20212021~2.4x (25M→60M)>$70B (Meta alone)Downsized, niche
Blockchain/Web32014–2022202130–60M users after 15 yrs$100B+ VCSurviving but niche
Chatbots "New Apps"2016201670% failure in 6 moModerateDead as "new apps"
IoT2010–2015201612–15B (vs 50B predicted)$100B+Enterprise yes, consumer meh
3D Printing2012–20142013Consumer: dead$604M (MakerBot)Industrial niche
Voice/Alexa Skills2014–20182017–2019160K skills, <1% active devs$10B/yr lossesKilled dev program
Smartphones2007–20102012–2014~5.5x (122M→680M)N/AUbiquitous
The Actual WebN/A1999–200025x (16M→400M)~$5T bubbleUbiquitous
The most dangerous precedent is 2016 chatbots, not VR.
  • Satya Nadella (March 2016): "Bots are the new apps." Almost identical to "agents are the new websites."
  • Facebook launched Messenger Bot Platform. 30,000 bots within months. 70% couldn't fulfill requests.31
  • By September 2016 — 5 months after launch — Facebook's VP admitted: "It got really overhyped really, really quickly."
  • The obvious counter: LLMs in 2025 are categorically better than 2016 NLP. True. But "the tech is better this time" is what every hype cycle says.

The Five Conditions for a Real "1995 Moment"

Analyzing what the web and smartphones shared — and what VR, blockchain, chatbots, IoT, 3D printing, and voice all lacked:31

ConditionAgentsScore
Zero-friction first contact (no hardware, no download)Text a prompt. Lower friction than the web (no URL needed).PASS
Day-1 utility (solves a real problem immediately)Research, writing, coding, customer service — immediately useful.PASS
Better than what it replaced (not just different)An agent doing your research IS categorically faster than Googling.PASS
Network effects that compoundUnclear. Individual agents don't inherently make other agents better. The web's network effect was content. What's the equivalent for agents?WEAK
Universal addressable marketEveryone delegates tasks. But "everyone needs an agent" is an assumption, not evidence.UNCERTAIN
The Alexa Skills warning is underrated. Amazon built 160,000 "agent skills." Nobody could find them. The developer program was killed in 2024 ($10B/yr in losses). The lesson: without discovery, proliferation creates a graveyard, not an ecosystem. This is the exact risk for agents.31

VI. The Definition Problem: Is "Agent" Even a Category?

Before arguing about timing, we need to ask: are we even measuring the right thing?

Gartner tested thousands of products marketed as "agentic AI" and found only ~130 that are genuinely agentic.32 The rest are repackaged chatbots, RPA with fresh paint, and workflow automation with an "agent" label. The term "agentwashing" (analogous to greenwashing) is now in mainstream use.

The Salesforce Rename

Salesforce didn't just add agents. They renamed everything:28

Old NameNew Name
Einstein / AI CloudSalesforce AI
Customer 360Agentforce 360
Sales CloudAgentforce Sales
Service CloudAgentforce Service
Commerce CloudAgentforce Commerce

If a Salesforce customer reports "we're using AI agents," they may just be using Sales Cloud with a new logo. The "$500M ARR" is ~3% of Salesforce total revenue, and analysts note adoption is "lagging" with "below-consensus revenue guidance."28

What this means for the thesis

If "agent" means anything from a renamed chatbot to genuine autonomous reasoning, then any survey reporting "72% agent adoption" is measuring how many companies use ANY AI feature. The number is technically true and substantively meaningless. The real question isn't "are companies using agents?" — it's "are companies deploying truly autonomous software that acts independently?" And the answer to that question, by Gartner's own count, is: almost nobody.


VII. Timing: The Email Adoption Curve

If the personal agent is identity technology, it follows the email curve, not the website curve. Each identity technology had a specific "Hotmail moment" — the inflection where it went from technical-users-only to everyone-can-do-this-in-2-minutes. That moment determines everything.

2022–2024 — ARPANET ERA (~1971–1989 FOR EMAIL)
ChatGPT launches. LLMs go mainstream. But LLMs are brains without hands. AutoGPT, BabyAGI, Operator all try and stumble. The CONCEPT is proven. The PRODUCT isn't ready.
Email parallel: email exists on ARPANET. Only researchers and military use it. Requires institutional access.
2025–2026 — PRE-HOTMAIL ← WE ARE HERE
OpenClaw at 180K stars. ~10–50K humans running persistent agents. AAIF founded (MCP, goose, AGENTS.md). Standards forming. Setup: 20–30 min for technical users. This is sendmail-era email — powerful, self-hosted, requires technical skill. The Hotmail moment hasn't happened yet.
Email parallel: ~1993–1995. CompuServe, AOL email. Growing but requires ISP/institution. Not free, not trivial.
2027–2029 — THE HOTMAIL MOMENT
Apple, Google, or Microsoft embeds a personal agent in the OS. 2-tap setup. Free tier. Your agent gets your context from your existing accounts. This is the inflection. When it's as easy to get a personal agent as to create a Gmail account, adoption goes exponential.
Email parallel: Hotmail (1996), Yahoo Mail (1997), Gmail (2004). Free, instant, zero-friction.
2029–2033 — THE NETWORK EFFECT KICKS IN
Your agent talks to their agent. Scheduling becomes agent-to-agent. Commerce becomes agent-to-agent. Not having an agent means people can't reach you efficiently. Social pressure compounds. The IETF agent identity drafts become standards.
Email parallel: 2000–2005. "What's your email?" becomes default. Not having one = unreachable.
2033–2037 — NEAR-UNIVERSAL (DEVELOPED MARKETS)
50%+ of adults in developed countries have a personal agent. The agent IS your primary digital interface. You don't email — your agent emails. You don't browse — your agent browses. Digital activity without an agent feels like business without email felt in 2010.
Email parallel: 2005–2010. 90%+ in developed markets. Infrastructure for everything.
Timing estimate: Hotmail moment ~2027–2029. Majority adoption ~2033–2037. This is consistent with the identity technology acceleration pattern (telephone 60 yrs, email 30, smartphone 8). A personal agent inherits the infrastructure of ALL previous layers (internet, cloud, smartphones, messaging). Each identity tech adopted faster than the last. But the trust barrier is genuinely novel — delegating autonomous action is harder than delegating communication. Add 2–3 years vs. the smartphone curve. The bear case (AI plateau, regulation) adds 3–5 more years. Range: 2033–2040.

VIII. Portfolio Reassessment (Honest Version)

The first version of this section was biased — force-fitting your existing projects into the macro thesis. Here's the honest reassessment with the bear case applied.

claw.degree

REAL GAP — POSSIBLY PREMATURE

The gap is real but the problem may not exist yet. PageRank was invented in 1998 when there were 2.4 million websites. Google didn't come before the proliferation — the proliferation came first, THEN the quality layer. With only ~130 genuine agents (Gartner), who is asking "should I trust this agent?" today? The answer: almost nobody yet. This could be 3–5 years early. The HubSpot Website Grader playbook worked because millions of websites already existed.

Agent Elo

MORE PREMATURE THAN CLAW.DEGREE

Ranking needs volume. Elo ratings require a population of contestants. If there are 130 genuine agents, there's nothing to rank. The Alexa Skills precedent haunts this: 160K "agents," nobody could find them, the developer program was killed. Discovery solves a problem after proliferation, not before. Even more timing-dependent than claw.degree.

avet — Agentic Vetting

LEAST TIMING-DEPENDENT

Community vetting is a current problem, not a future one. Bot spam, fake accounts, and low-quality members exist today — independent of whether "agents" proliferate. avet solves a problem that exists NOW (community quality gating) and gets more valuable if agents proliferate. Least dependent on the macro thesis being right.

OpenClaw Hosting

WEAKENED

Hosting commoditizes fast. E2B ($32M), AWS AgentCore, Cloudflare Workers all entering. The 2016 chatbot parallel applies: early infra players got wiped when the platforms absorbed the functionality. Manual concierge still works for cash flow but the platform dream is weak.

Blue-Collar AI (Wenhao)

ACTUALLY STRENGTHENED BY THE BEAR CASE

Vertical > horizontal in a hype cycle. If the bear case is right (agents overhyped, most fail, adoption slower), then narrow vertical applications with real customers beat horizontal infrastructure plays. Wenhao's packing QC doesn't depend on "agent" as a category existing. It depends on computer vision working for a specific task. The least thesis-dependent play in the portfolio.

The "Trust Stack" Narrative

SEDUCTIVE BUT PREMATURE

I was building a narrative, not analyzing data. Framing claw.degree + Agent Elo + avet as a "trust stack for the agent economy" sounds compelling but requires: (1) agents proliferate enough to need trust scoring, (2) users choose agents independently (not via SaaS embedding), (3) the discovery problem is real. None of these are confirmed yet. The trust layer is real but it's a 2030 opportunity, not a 2026 one.

The deeper insight: focus on what's true TODAY, not what might be true in 2030.
  • Blue-Collar AI solves a problem that exists today (packing QC, safety inspections). Ship it.
  • avet solves a problem that exists today (community spam/quality). Ship it.
  • claw.degree solves a problem that will exist. Build the MVP as a lead-gen tool (like HubSpot Grader) but don't bet the farm on it.
  • Agent Elo depends on a world that doesn't exist yet. Keep the domain. Don't build until there are agents to rank.
The macro thesis can be right AND your timing can be wrong. Netscape was right about the web and still went bankrupt.

IX. Full Red Team: Claim-by-Claim Challenge

Original ClaimCounter-EvidenceRevised Assessment
"72% of enterprises using/testing agents" Census Bureau: 5–14% use ANY AI. Vendor surveys use self-selected samples. "Testing" includes "tried ChatGPT once."25 MISLEADING Real adoption is 5–14%. Use Census data.
"We're at late 1996 / early 1997" Gartner: agents at Peak of Inflated Expectations. GenAI in Trough. 95% of pilots fail. ChatGPT = Mosaic (1993), not Netscape (1995).32 REVISE TO ~1995 Post-awareness, pre-commercial.
"Agentforce $500M ARR proves enterprise demand" Salesforce renamed entire product line. 3% of total revenue. Analysts call adoption "lagging." Multi-step success: 35%.28 INFLATED Rebrand revenue ≠ agent revenue.
"5–7 years to 'every business has agents'" Cloud computing took 10 years (2006→2016). Self-driving promised by 2020, still not here. Technology adoption is almost always slower than predicted.33 REVISE TO 8–15 YRS 2032–2037, not 2028–2030.
"Trust layer is wide open — build now" PageRank emerged after 2.4M websites, not before. Only 130 genuine agents exist. Who needs trust scoring for 130 products? DIRECTIONALLY RIGHT, 3–5 YRS EARLY
"Agent = coherent category" No standard definition. Gartner: 130 real agents out of thousands. Forbes: "agentwashing." Salesforce/Microsoft rebranded existing products.32 CATEGORY NOT YET DEFINED
"1995 framing understates the speed" 7 technologies got this baptism. 5 failed. VR: $70B burned. Chatbots: 70% failure in 6 months. Alexa: $10B/yr losses.31 AGENTS PASS 4 OF 5 CONDITIONS Stronger than VR/crypto but network effect is unproven.
"80% report measurable ROI" PwC 2026 CEO Survey (4,400 CEOs): 56% report ZERO financial benefit from AI. MIT: 95% deliver zero P&L impact.2726 CONTRADICTED BY BETTER DATA
"Models keep improving" GPT-5 was a "botched non-upgrade." MIT Tech Review: "the great AI hype correction of 2025." Hallucination unsolved. Best agent: 24% task completion.2934 PROGRESS IS REAL BUT PLATEAUING
Gary Marcus (AI researcher, NYU) — the best steel-man of the bear case: "AI Agents will be endlessly hyped throughout 2025 but far from reliable, except possibly in very narrow use cases." By Aug 2025, he assessed this as confirmed. His core argument: hallucination is unsolved, and without solving it, reliable agents are impossible. He believes agents will eventually be worth trillions — but "eventually" might be years or decades, not months.34

X. Verdict

The personal agent — as distinct from the LLM, as distinct from the enterprise chatbot rebrand — is identity technology. Identity technology converges on universality. This is the thesis.

The publishable insight: There's a testable taxonomy that predicts which technologies become universal and which plateau. Identity technologies (telephone, email, smartphone, social media) ALL reached near-universality. Tool technologies (VR, 3D printing, IoT) ALL plateaued. The distinguishing properties are measurable: network effects, social pressure, infrastructure dependency, cost trajectory. A personal agent in the OpenClaw/Donna sense — persistent, contextual, operating your digital life, representing you — has all four identity properties. It follows the email curve, not the VR curve.

Where we are: Pre-Hotmail. ~10–50K humans running persistent personal agents globally. OpenClaw at 180K stars, ~1,100 verified deployments. This is ARPANET-era email (~1985) or the web at 130 sites (1993). The "Hotmail moment" — Apple/Google/Microsoft embedding a personal agent in the OS with 2-tap setup — is ~2027–2029. After that, the identity adoption curve kicks in.

Why the generic "AI agent" bear case doesn't kill this thesis: The red team (Section IX) is devastating for the generic category: 5–14% real adoption, 130 genuine agents, 95% pilot failure, agentwashing everywhere. All true. All measuring the wrong thing. The generic category includes Salesforce renaming Sales Cloud. That's not what this thesis is about. This thesis is about a persistent entity with your context, your channels, and your autonomy. The bear case for THAT is different: the trust barrier is genuinely novel (delegating autonomous action ≠ delegating communication), and the Hotmail moment hasn't happened yet. But the direction is not in question.

Timing: Hotmail moment ~2027–2029. Majority adoption ~2033–2037. Bear case adds 3–5 years (2036–2040). The identity acceleration pattern (telephone 60 yrs, email 30, smartphone 8) suggests agents inherit compounding infrastructure advantages. But trust delegation is harder than communication — add time for that.

For building now: You're running the ARPANET-era equivalent. Donna is the reference implementation. The 10–50K early adopter community is real and growing. Build for THEM — the people who already run persistent agents. claw.degree scores agents for this community (not mass market). avet gates communities against agent spam (a problem that exists today). Blue-Collar AI solves today's problem with today's technology. Agent Elo waits for volume. The generic "AI agent" hype cycle is noise. The personal agent adoption curve is signal. Ignore the noise. Follow the signal.


References

[1] IDC FutureScape — 1 billion agents by 2029. Aspirational, not descriptive.
[2] Zapier Survey 2025 — 72% using/testing. CAUTION: n=~500, self-selected. Contradicted by Census [25].
[3] MCP Protocol — 97M+ SDK downloads/month. Linux Foundation Dec 2025.
[4] Internet Live Stats — Website count 1991–2000.
[5] Internet History Podcast — Netscape IPO.
[6] ICANN / Network Solutions — NSI revenue $5M→$93.7M.
[7] Precedence Research — AI agents $7.9B (2025), $236B (2034).
[8] SearchYour.AI — 80% ROI. Vendor survey. Contradicted by [27].
[9] Gartner — 40% of apps embed agents by 2026. 40%+ projects canceled by 2027.
[10] Salesforce Q3 FY26 — $500M+ ARR. Includes rebranded products [28].
[11] Crunchbase / TechCrunch — $202B AI VC, Cognition $10.2B/$73M ARR.
[12] Google A2A — 100+ partners, RC v1.0.
[13] OpenAI Agents SDK (March 2025).
[14] E2B — $32M, 88% of Fortune 100.
[15] x402 Protocol — 75.4M txns, Coinbase-backed.
[16] Braintrust / observability ecosystem
[17] IETF Agent Identity Drafts — draft-yl-agent-id-requirements-00, draft-rosenberg-ai-protocols-00. When IETF drafts identity specs, they expect universality. Same stage as RFC 733 (1977) for email.
[18] UnmitigatedRisk.com — Agent timeline prediction.
[19] Dotcom Bubble
[20] GeoCities — 3.5M users, $3.9B, never profitable.
[21] Felicis — "The Agentic Web"
[22] MIT Web Growth Summary
[23] Bill Gates — "Everyone will have an AI-powered personal assistant within 5 years." Fortune, Nov 2023.
[24] Protiviti
[25] U.S. Census Bureau BTOS — 5.4% use any AI. Peaked ~14%, then declined. Gold standard. 1.2M firms.
[26] MIT NANDA / RAND — 95% deliver zero P&L. 80% fail in 6 months.
[27] PwC 2026 CEO Survey — 56% zero benefit. 4,400 CEOs, 95 countries.
[28] Reuters / NoJitter / Coastal Cloud — Agentforce rebranded. 35% multi-step success.
[29] Carnegie Mellon — Best agent: 24% task completion.
[30] S&P Global — 42% abandoned AI initiatives by mid-2025.
[31] "1995 for X" graveyard — VR, blockchain, chatbots, IoT, 3D printing, voice. Full bibliography: crm/notes/red_team_agents_websites_thesis.md.
[32] Gartner Hype Cycle 2025 — Agents at Peak. ~130 genuine. 40%+ canceled by 2027.
[33] Adoption timelines — Cloud: 10 yrs. Self-driving: still not here. Amara's Law.
[34] Gary Marcus — "Overhyped. Worth trillions — eventually."
[35] METR — AI tools made devs 19% slower. 39-point perception gap.
[36] Personal agent adoption research — OpenClaw: 180K stars, ~1,100 verified deploys (Shodan/Censys), 8,900 Discord. AutoGPT: 140K stars. Open Interpreter: 62K. Goose: 30K. Humane AI Pin: dead ($200M). Rabbit R1: 5K DAUs. OpenAI Operator: killed after 7 months, folded into ChatGPT. DigitalOcean, Sterlites, AIHola, TechRadar, Wikipedia.
[37] Identity vs. tool adoption data — Telephone: US Census historical tables. Email: Pew Research, Radicati Group. Smartphone: Statista, Pew. Social media: DataReportal, Our World in Data. VR: Security.org. IoT: ConsumerAffairs. 3D printing: GII Research. Full data tables: crm/notes/eric_agent_as_identity_thesis.md.
[38] Anthropic Economic Index — Automation 27%→39% of Claude usage, surpassing simple assistance. Session duration doubled (6→15 min).
[39] HBS / Perplexity Agent Study — 55% of agentic queries are personal use. 60% under 35. Top: productivity & research (57%). Hundreds of millions of interactions analyzed.
[40] Linux Foundation AAIF — Founded Dec 2025 by Anthropic, Block, OpenAI. Projects: MCP, goose, AGENTS.md. Validates agent layer as distinct infrastructure from LLM layer.
[41] Sam Altman — Personal agents as "AI's killer function." MIT Technology Review, May 2024. Satya Nadella — "An AI companion for everyone," separate work/personal agents. Microsoft Blog, Oct 2024. Goldman Sachs — "Rise of the personal agent" as key 2026 trend.