Skip to main content
← Daily Digest

Sunday, October 25, 2026

14 signals
10

The Man Who Led 100+ CROs (Even OpenAI's) | John McMahon - 5x Post-IPO CRO / The GOAT

Topline · GTM Ops · Practitioner Story · Oct 25
  • Hire on characteristics (intelligence, drive, coachability, adaptability, integrity) over resume credentials—character flaws are the primary reason hires fail, not skill gaps
  • AI is a research tool that amplifies existing rep quality: great reps get better, lazy reps get lazier; conversational intelligence tools only verify what was said, not whether true deal qualification occurred
  • First-time CROs must act as CEO peers with fiduciary responsibility—speak up with specific business critiques and solutions, not excuses; readiness for CRO role requires ability to recruit, coach, and lead people who want to follow
  • The PTC sales leadership lineage (100+ CROs including OpenAI's Dali Rajic) demonstrates that sales methodology and team-building principles are durable across market cycles and AI disruption
  • MEDDIC methodology remains relevant despite larger buying groups; the fundamentals of identifying champions, coaches, and enemies haven't changed with technology
10

Before you send the welcome email, do this first.

The Customer Success Café Newsletter · GTM Ops · Practitioner Story · Oct 25
  • CSM turnover is systemic: 1-in-7 companies repost identical roles within 7 weeks, indicating seat instability affecting customer continuity
  • Green accounts are deceptively high-risk: silence signals customer disengagement, not health; inherited accounts near renewal with no recent attention are churn vectors
  • Critical handover gaps exist in three areas: (1) verbal promises don't transfer to new CSMs, (2) renewal context/history lives outside CRM fields, (3) relationship trust resets to zero with each transition
  • Triage-by-noise creates renewal risk: teams prioritize escalating accounts while silent inherited accounts get welcome emails and queue delays, compounding neglect
  • Timing creates cliff risk: accounts inherited 90 days before renewal have insufficient time for relationship rebuild; accounts inherited 1+ year out have recovery window
10

Half Your CRM Data Is Wrong: AI Prospecting in a World of 49% Exec Turnover

SaaStr — Jason Lemkin · AI×GTM · Practitioner Story · Oct 25
  • First-party CRM data (past meeting transcripts, internal conversations) outperforms third-party vendor data for personalization because it contains unique context competitors can't access—demonstrated by the fish tank example that booked a meeting third-party data couldn't enable
  • 49% executive turnover in B2B GTM roles creates both a problem (stale contacts) and an opportunity—AI agents can track job changes via LinkedIn/press and reach prospects at new companies while also identifying replacement contacts at old companies.
  • Generic AI SDR sequences fail because they lack guardrails and human judgment; Aurasell's approach uses dynamic replanning (changing sequence length and channel based on engagement signals) plus mandatory human review before sends, preventing the self-replying-to-itself failure m
  • Email is a weak channel for sales leaders (60% of Aurasell's meetings come from cold calls, 40% from LinkedIn, almost none from email) because buyers are oversaturated; AI agents must be channel-aware and respect deliverability constraints across multiple mailboxes.
  • Human skills remain irreplaceable—SDR candidates are tested on hand-written prospecting emails and cold calling ability because AI handles research/sequencing/follow-up, but creative pitch control and relationship-building still require humans; this creates a hiring gap as AI-nat
9

Jason’s Takes on This Week’s 20VC: Meet 100 VCs, None-and-Done Beats One-and-Done, and Build-to-Prod Is Where I’d Make 10 BetsTime-Sensitive

SaaStr — Jason Lemkin · GTM Ops · Thought Leadership · Oct 25
  • Capital concentration is real: top-decile growth outside AI infra requires 100 VC meetings, not 15. Plan for 2-3 months and treat mega-funds as lottery tickets, not the plan.
  • Fund math matters more than pitch quality: mega-funds need $25B+ outcomes, not $2B. Founders must divide fund size by 20 and match realistic outcomes to fund stage before pitching.
  • None-and-done beats one-and-done in AI era: with AI reducing team requirements to $10M ARR, bootstrapping is now viable for more founders. One VC on cap table creates exit pressure without exit economics.
  • AI revenue is unstable: model 20-30% annual churn risk on usage-based AI products. Replit's routing flexibility (30-40% COGS swings) proves open-weight quality hasn't displaced incumbents in production.
  • Build-to-prod is the real bottleneck: code generation is solved; the 90% of time spent reviewing, testing, and hardening code is where 10 bets should be made. Claude Code Review validates the category.
9

An 86-minute guide to finding PMF | Rob Snyder (Author, The Power of Pull)

Lenny's Newsletter · GTM Ops · Deep Dive · Oct 25
  • The PULL framework (Project, Unavoidable, List of options, Limitations) provides a structured alternative to design-partner-heavy approaches for finding PMF
  • Rob's contrarian stance against design partners challenges conventional startup wisdom and suggests cold outreach + first sales calls are more revealing
  • The 'goblin mode' concept—what to do when nothing works—addresses the psychological/tactical crisis point most founders face but rarely discuss
  • PMF discovery is often accidental rather than methodical; two years of wrong direction before breakthrough suggests founders should expect non-linear paths
  • Framework applies beyond startups to innovation inside larger companies, expanding relevance beyond early-stage audience
8

Your Slack, Zoom, and call transcripts are probably your most valuable data. Few companies use them safely.

revops · GTM Ops · Quick Take · Oct 26
  • Conversation data (Slack, Zoom, call transcripts) represents underutilized competitive intelligence for RevOps teams
  • Data safety/governance gap exists: most companies lack frameworks to extract value while maintaining compliance and privacy
  • Emerging tension between data monetization potential and regulatory/ethical constraints in conversation intelligence space
8

Your Top Job For 2027: Hire 1-2 Execs Who Can Take the Load, Not Just Do the Work

SaaStr — Jason Lemkin · GTM Ops · Thought Leadership · Oct 25
  • The critical distinction: 'work' (tasks/projects) vs. 'load' (accountability/worry that stays with you). Most executives take on work but leave load with the founder—this is the real hiring problem.
  • Ex-founders with intact capacity are ideal load-carriers because they've owned P&Ls, made hard calls, and think like owners. But many ex-founders are still running on empty 5-10+ years post-exit and can only advise, not carry.
  • Six concrete signals that an exec is actually carrying load: you stop checking their work, they make 51/49 decisions independently, they push back on you, they handle crises without you, your board trusts them directly, and you can take 2 weeks off without everything breaking.
  • Founder burnout is real and persistent—Lemkin's candid admission that he's still 'a bit burnt' 10+ years after EchoSign exit, despite great team and successful exit, validates that this isn't a weakness but a structural consequence of founder-level responsibility.
  • Actionable 2027 priority: Hire 1-2 execs who can carry load BEFORE the next product launch or sales hire. Then actually hand them the decisions and accountability—founders often hire for load-carrying but never delegate the real weight.
7

OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?Time-Sensitive

Max Woolf's Blog · AI Research · Deep Dive · Oct 25
  • LLM pricing has collapsed 57% (to 43.3% of April 2026 cost) in 4 months due to competitive pressure from open-weight models and 29+ inference providers on OpenRouter, not IPO optics or loss-leading strategy
  • Cache read costs are the real battleground for agentic AI: 96-99% of token spend in coding agents, making a 75% cache reduction equivalent to 70-75% total cost savings—this is where differentiation happens
  • Open-weight models (DeepSeek V4.1 Flash, Kimi K3.1 Pro) are catching up fast on quality while maintaining 50% lower pricing, forcing OpenAI/Anthropic into a margin-compression race rather than a pricing power game
  • The 'token price will spike after IPO' narrative is structurally unsound: with low switching costs, negligible barriers to inference provider entry, and 29 competing hosts, prices will trend toward marginal cost, not artificial scarcity
  • Both OpenAI and Anthropic are pivoting moat strategy from pricing power to enterprise commitments and personal agents (OpenAI Dots, Claude Duet) because commodity inference margins are being competed away
6

AI delivers ‘zero financial payoff’ to most companies, Apollo economist saysTime-Sensitive

AI | TechCrunch · Enterprise AI · Quick Take · Oct 25
  • Macro-level data shows AI has delivered zero measurable financial benefit to non-tech Fortune 500 companies despite massive investment—profit margins and revenue-per-worker remain flat
  • The gap isn't a timing issue; Slok argues structural: AI lacks judgment, accountability, trust, cultural understanding, and ethical reasoning—the actual drivers of business value and revenue growth
  • This contrarian thesis directly challenges the AI-SDR and AI-sales-tool narrative; suggests companies may be investing in productivity theater rather than actual ROI-generating technology
  • Emerging counter-narrative to AI hype cycle: validates skepticism about enterprise AI adoption and ROI claims from vendors
6

Surfe vs Lusha (2026): accuracy, data sources, credits and AI access compared

Lusha's Blog - B2B | Sales | Marketing | Recruiters | News · AI×GTM · Vendor Content · Oct 25
  • Find rate and accuracy are fundamentally different metrics—Lusha publishes accuracy (98% email on verified records), Surfe publishes find rate (85%)—making direct comparison misleading without understanding methodology
  • Data decay is measurable and material: 12.25% of US sales leaders change roles annually (~1% monthly), requiring quarterly re-verification of leadership lists regardless of tool choice
  • Credit models differ structurally: Lusha uses one shared balance across all channels (1 credit/email, 5/phone), while Surfe separates email and mobile find credits, charged only on validated results—cost comparison requires calculating 'cost per usable contact' not per credit
  • Lusha emphasizes dated signals and transparency (26 signal types with effective dates, last-update timestamps on every record), while Surfe emphasizes waterfall breadth (15+ providers) and real-time lookups—different philosophies on freshness vs. depth
  • Coverage differs by unit: Lusha reports 515M verified contacts across 56 markets; Surfe reports 1B+ profiles—comparing on target list by region/seniority matters more than headline totals
5

Podcast: Kimi K3.1 Pro, the new cyber reality, and Muse musingsTime-Sensitive

Interconnects AI · AI Research · Deep Dive · Oct 25
  • Kimi K3.1 Pro shows benchmark inconsistency: top-tier on Artificial Analysis but 20 points behind on Terminal-Bench Science, suggesting RL training optimized for short-horizon exploitation over exploration
  • Three trillion parameter models face infrastructure constraints (require B300 nodes) that may limit adoption compared to smaller Flash-style alternatives, creating potential deployment friction
  • Responsible open-weight release strategy (deliberately downweighting offensive cyber capabilities while maintaining defensive) positions Kimi against fearmongering narratives and increases scrutiny legitimacy versus closed models
  • Cybersecurity professionals distinguish between offensive (long-horizon, persistence-requiring) and defensive (vulnerability detection) capabilities—open models sufficient for latter, closed models superior for former
  • Model launch professionalization accelerating: from uncoordinated drops to coordinated inference framework availability, exclusive access windows, and (sometimes undisclosed) paid promotion campaigns
5

Apollo analysis shows no one besides the biggest tech giants is seeing benefits from AITime-Sensitive

SiliconANGLE · AI Market · Quick Take · Oct 26
  • AI ROI is concentrated in Magnificent Seven tech giants; S&P 493 companies show zero productivity gains despite massive investment since 2019
  • Enterprise spending behavior is shifting: AI viewed as cost center, not growth driver; mainstream companies unwilling to pay for latest models
  • AI's core limitation in business context is judgment-based decision-making requiring context, culture, ethics, and trust—areas where AI still struggles significantly
  • The AI narrative will be decided by non-tech corporate America's ability to monetize AI; current evidence suggests this transformation is harder than predicted
  • Venture-backed startups increasing AI spend while established firms pull back—creating a bifurcated market with limited macroeconomic impact
5

Elon Musk predicts neocloud shakeout as credit for AI infrastructure dries upTime-Sensitive

SiliconANGLE · AI Market · Quick Take · Oct 26
  • Credit markets have bifurcated: only companies with diversified revenue (Starlink, launches) can access AI infrastructure debt; pure-play neoclouds face funding drought
  • Predicted 200+ neocloud consolidation wave driven by tightening lending standards and investor skepticism about GPU ROI timelines
  • Hyperscaler capex arms race ($1T+ in 2027) is sustainable only for AWS/Microsoft/Google/Meta; smaller competitors face acquisition or failure
  • Nvidia demand risk: neocloud failures could reduce GPU demand, creating revenue pressure on chip suppliers despite hyperscaler spending
  • Credit market stress signals: investors demanding higher yields on AI infrastructure bonds due to concerns about revenue-to-capex ratio sustainability
4

Van Jones’ advice for bridging political divides at work

Charter - Future of Work, AI, Management, Hybrid · Future of Work · Thought Leadership · Oct 25