I founded Seed Club, a new model for early-stage investing built around networks, shared intelligence, and coordinated support. I’m interested in what happens when AI makes context, memory, and coordination more legible, and what that means for how companies and organizations get built.
My agent drafts this site from what I save. Green is me.
Agents reshape dev workflows but not the need for comprehension 6 signals ▾
New coding workflow probes known and unknown gaps in codebases · @dzhng Sierra says internal AI agents became core to running the company · @noahhorner_ Product specs replace PRDs in agent driven software teams · @gokulr Developers still need to understand code written by agents · @kieranklaassen Agent loop turns user interviews into paying software · @pzakin Developers still need to understand code written by agents · @geoffreylitt
AI coding agents are replacing traditional specs with leaner formats and automating codebase discovery, yet a strong counter-current holds that developers who cannot read and understand agent-generated code will lose the ability to catch errors and maintain systems.
- Product specs are replacing PRDs as the primary unit of product work
- Internal AI tools now handle both execution and institutional memory at fast-growing companies
- Scanning for unknown unknowns in a codebase is becoming a standard workflow step
My take · Jess · Jul 4
While many fear what ai will mean for startups and software products, I'm a big believer that the firm is becoming software, and startups have an advantage at first principles rethinking how entire industries are serviced with ai.
Enterprise AI trust and security concerns grow 4 signals ▾
Anthropic restores Fable 5 with tighter cyber safeguards · @alexstamos Security concerns trigger enterprise ban on Claude Code · @__tinygrad__ Alleged client-side checks raise trust concerns around coding tools · @AISuperDomain Poor AI product design can make outputs impossible to verify · @HamelHusain
Anthropic's developer tools face escalating enterprise trust problems, with unverified spyware allegations and claimed geo-targeted surveillance checks in Claude Code driving bans, while questions about whether AI outputs can be reliably verified at all deepen the credibility gap.
- Alleged timezone checks targeting Chinese users have prompted enterprise-wide bans
- Fable 5 jailbreaks reportedly added no novel capabilities beyond existing models
- Hard-to-verify AI output signals a product design failure, not just a technical limit
My take · Jess · Jul 3
While interest in this seems to be picking up, Im surprised by how little concern enterprises have over protecting their data and workflows from model providers or integrators. I expect this trend of increased caution to continue.
Frontier AI moat claims are fragmenting under pressure 8 signals ▾
Venture capital built the wrong AI and pension savers may pay · @AndrewOrlowski AI vendors push conflicting moats across models, tooling and tokens · @0interestrates Benchmark gains do not prove weaker models match frontier AI · @alexframegreen Claims that AI will be winner take all may not hold · @sebkrier Open models gain ground as trust in frontier labs frays · @gbrl_dick AI capability gains still look early, not near commoditization · @deanwball Open source model costs are forcing enterprises to rethink AI security · @quxiaoyin AI progress may depend on economic capability, not tokens · @lefttailguy
The winner-take-all narrative around frontier models is under strain as open-source alternatives close capability gaps and enterprise trust in closed labs erodes, while competing moat claims across model providers, harness builders, and token sellers undercut any single dominant position.
- Benchmark gains on weaker models do not replicate frontier output quality
- Open-source hosting lets enterprises sidestep data-sharing concerns entirely
- Cost pressure from open models is forcing a rethink of AI security assumptions
Venture pricing detaches from fundamentals, echoing 2021 7 signals ▾
Angel investing works on belief and affinity, not VC logic · @jayyeh Early angel rounds work best when founders start close to home · @christophersaum Top venture investors warn the funding boom looks like 2021 · @madhavchanchani VC megafund fee terms can erase gains for LPs · @credistick Founders should build an angel base before pitching VCs · @jayyeh Fraud and ghost shares taint startup secondary investing · @rsg European founders struggle to raise from Bay Area VCs · @arian_ghashghai
Senior partners at prominent funds are warning that valuations have become untethered from reality in a pattern matching 2021, while megafund fee structures, SPV fraud in secondaries, and geographic bias compound the disadvantage for limited partners and non-Bay Area founders.
- Angels function as credibility signals and intro nodes, not only capital sources
- Megafund fee drag can eliminate hundreds of millions in LP returns versus PE peers
- Secondary market fraud and ghost shares go underreported because victims find it embarrassing
Token and equity alignment stays structurally unresolved in crypto 6 signals ▾
Robinhood 7 percent DeFi yield relies heavily on incentives · @HorinekPM Robinhood DeFi launch ties a rival product to Coinbase rails · @sealaunch_ Venice token thesis leans on burns, profits and agent markets · @nikshepsvn Venice offers a rare case for a true AI utility token · @DrNickA Crypto token structures hinge on trust and alignment · @knimkar Crypto projects may still need both tokens and equity · @lex_node
As projects raise equity alongside tokens, the question of whether token holders and shareholders have aligned incentives remains contested, with protocol-level burn mechanics earning the most trust and dual structures drawing sustained skepticism.
- Long-duration warrants on tokens signal genuine institutional conviction
- Robinhood's DeFi yield mixes native lending with incentive campaigns on a rival's rails
- Utility tokens with on-chain revenue visibility offer a rare verifiable case
Latest signals · Monday, July 6 all signals →
new What software companies can't measure is what keeps them running · seangoedecke.com new Why states fail when they simplify what they can't fully see · ribbonfarm.com Earlier weeks
AI economic value migrates below the model layer
The competitive question has shifted from model capability to where margin actually concentrates. Workflow orchestration, enterprise flywheels, and edge distribution are all contending as the real prize, while model providers face commoditization pressure from multiple directions.
- Embedded workflow lock-in may outlast any model-level advantage
- Enterprise flywheels capturing tacit knowledge could prove more durable than model leads
- Edge providers capturing economic value have strong incentive to keep it private
Applied AI value may accrue to workflow orchestration layers · @i_zerbib Applied AI value may accrue to workflow orchestration layers · @levie Enterprise AI moats may come from embedded workflow lock in · @alexolegimas Enterprise AI shifts toward proprietary model flywheels · @AravSrinivas AI stack specialization could mirror cloud's giant winners · @masonnystrom Frontier AI labs face pricing pressure at the edge · @hypersoren Enterprise AI may spread faster through chat than seat licenses · @_simonsmith AI startup moats are diverging across labs data and distribution · @aashaysanghvi_
Frontier model access shifts to government approval gates
Government approval gates emerged for GPT-5.6 access on security grounds, as a massive distillation attack on Claude attributed to Alibaba confirmed that capability extraction is a live threat. Federal AI policy simultaneously hardened toward restricting Chinese frontier model access globally.
- Gated rollout creates a two-tier market of approved and non-approved users
- Distillation attacks may become the primary vector for capability transfer between rivals
- Regulatory trajectory mirrors the protocol-level battles crypto already fought
Security restrictions could shut Chinese frontier models out globally. · @teortaxesTex Federal AI policy shifts toward tighter and murkier control · @deanwball OpenAI unveils GPT-5.6 lineup with three model tiers · @OpenAI AI labs face the same regulatory battles crypto already fought · @ccatalini U.S. moves to approve GPT-5.6 access customer by customer · @ns123abc OpenAI asked to stagger GPT-5.6 launch over security fears · @steph_palazzolo Anthropic says Claude faced a massive model distillation attack · @MTSlive
Agents split enterprise AI into two distinct problems
Building with agents divides into two fundamentally different challenges: restructuring internal operations versus rebuilding products as agents, and conflating the two is a key failure mode. Internally, agent adoption is already visibly compressing white-collar roles and dissolving traditional team structures at leading firms.
- Role boundaries between engineering, product, and design are dissolving
- Sharing successful agentic workflows across a team remains unsolved
- Executives treating AI as a compliance checkbox are misreading the transition
Agent adoption splits between internal operations and products · @ankrgyl AI teams may reorganize around product lifecycle archetypes · @bcherny Executives warned AI will remake companies, not tick a box · @Steve_Yegge Teams still lack good ways to share agent workflows · @dwr Teams still lack good ways to share agent workflows · @dwr Agentic workflows could compress white collar labor demand · @_The_Prophet__ Meta product teams replace process with internal AI agents · @nikhyl
Moats thin as mega fund capital distorts every stage
Conventional defensibility is weakening to where founder reputation has become the primary moat signal, with teams outweighing decks in early-stage evaluation. Mega fund over-capitalization is simultaneously lowering bars across all stages, inflating rounds throughout the funding stack rather than specifically targeting seed.
- Sitting out a bubble may carry more career risk than joining it
- Seed fund strategy is shifting toward diversification and option value
- A large liquidity wave may be approaching across major private companies
AI startup moats may come from multiplayer systems, not models alone · @illscience Startup moats weaken as investors fixate on teams · @zamdoteth Seed funds shift toward broad portfolios and option value · @masonnystrom Venture firms brace for a new wave of startup liquidity · @Samirkaji New research examines how active mega funds are in seed · @pavelprata Venture incentives can reward joining bubbles over sitting out · @TurnerNovak Mega funds may be inflating startup rounds across every stage · @gdibner Venture capital may split between swarms and permanent capital · @joeljohn
Open models are closing the frontier performance gap
As capable open models approach near-parity with closed frontier models, the cost differential is becoming decisive. Frontier labs have priced for margin rather than compute cost alone, and that margin looks fragile when open alternatives offer roughly 90% lower cost at a 10% performance penalty.
- Subsidized inference may have been masking open model viability all along.
- Token cost pressure at scale is pushing legal AI toward post-training open models.
- Capital allocated to frontier AI could be mispriced if intelligence commoditizes.
Frontier model pricing reflects quality, margin and scale economics · @tugot17 Harvey turns to open models as token costs crush margins · @willchen500 GLM 5.2 could ignite a lasting boom in rented GPU demand · @goodalexander AI teams face a brutal new era of token ROI and cost control · @adambcohen93 GLM-5.2 could be the first open model to replace closed ones · @gneubig AI capex may overshoot demand and trigger a boom-bust cycle · @orrdavid Open source AI progress could push governments toward censorship · @goodalexander A 10% intelligence gap at 90% lower cost with $5 trillion at stake · @zerohedge The chart that breaks the AI trade when intelligence gets cheap · @_The_Prophet__
Multi-agent orchestration reframes what a model is
The Sakana Fugu launch illustrates a structural shift where frontier-level performance is achieved by orchestrating swarms of smaller models rather than scaling a single one. Intelligence is beginning to look less like a product and more like a supply chain.
- Codex-style loops can audit, test, and fix entire codebases autonomously at scale.
- Effective loops require real-browser verification and environment tooling, not just prompts.
- The single-model API abstraction increasingly conceals a multi-agent system beneath.
Codex loops can audit, test, and fix apps at massive scale · @tomosman Builders are swapping examples of Codex and Claude loops · @argofowl Sakana launches Fugu to match frontier AI through orchestration · @SakanaAILabs AI orchestration turns models into a commodity supply chain · @_The_Prophet__ Agent loops get stronger with better testing and tooling · @ctatedev Sakana Fugu runs a learned multi-agent swarm behind one model API · sakana.ai A builder names the three-tool AI agent stack as his best workflow · @yitong
Venture capital concentration is distorting deal quality
Capital has consolidated into larger funds whose fee structures push toward consensus bets and narrative-aligned investments, leaving genuine outliers chronically underfunded. Seed is effectively two markets, with roughly 10% of deals capturing half the dollars and most headlines.
- Emerging managers lost ground during the ZIRP boom, not after it.
- Fee incentives drive GPs toward scale and safety rather than discovery.
- Founders can resist valuation pressure by forcing investors to name a counter-number.
Founders should force investors to make real pricing offers. · @amiklas Fund economics push venture investors toward safer consensus bets · @gdibner A small hype segment dominates seed capital and media attention. · @tmrohan Venture is consolidating into larger funds over diversification · @credistick Capital concentration has distorted venture more than most admit · @Rick_Zullo Early-stage venture is having a standout year as a category · @adityaag Emerging managers lost ground through ZIRP, not the reverse · @credistick
AI is splitting engineering teams and repricing software
Generative coding tools are fracturing engineering into those who generate and those who review, creating a visible class divide and a profession-wide identity crisis. Simultaneously, software pricing is shifting away from seat licenses toward alignment with actual business outcomes.
- The 'context layer,' infrastructure making AI useful for real codebases, is a major emerging VC thesis.
- Coding agents remain too 'software-brained' to generalize cleanly to broader knowledge work.
- Taking humans fully out of high-stakes loops remains a 'brutally long slog' even with mature tooling.
Coding agents still struggle to fit real knowledge work · @emollick Startups race to build the context layer for engineering AI · @GergelyOrosz AI coding is splitting engineering teams into pushers and reviewers · @deedydas AI is pushing software from seat sales toward paid business outcomes · @rodriscoll Taking humans out of the loop is brutally hard but worth it. · @Jason Howie hit 50% autopilot after years of human-in-loop scheduling · @awwstn
Intent-centric OS and data firewalls reshape platform power
Apple's announced shift toward intent-centric interfaces that construct UI on demand threatens the app ecosystem model. Platforms blocking LLM data access are spawning decentralized crawling marketplaces as a countermeasure.
- On-demand interface generation could eliminate third-party apps entirely
- Permissionless data markets emerge wherever platforms erect LLM firewalls
AI agent costs exploded, forcing outcome-based pricing models
The shift from chat to agents drove costs far beyond early estimates, as recursive agent spawning multiplied token consumption. Firms that price for outcomes over token volume can capture labor-scale revenue, but only by treating every token and hour as margin.
- Token spend at scale blurs the line between software and services
- Open question: which firms can actually measure outcome quality reliably
Real autonomy is finding work, not executing tasks
The defining feature of a truly autonomous loop is not task execution but the ability to propose work without a human prompt. Self-improving systems that search proactively over potential improvements against stated objectives represent the next capability threshold.
- Proactive explorer agents differ fundamentally from reactive task agents
- Self-improvement applies at the organizational level, not just software
Individual AI assistants give way to team orchestration
The market needs an orchestration layer spanning an entire team's workflow, combining work management, agent assignment, and shared context. Codified, programmable workflows are replacing knowledge-based skills as the reliable unit of agent execution.
- Shared MCP servers and context compound value across teams
- Codified workflows require less model intelligence, enabling more reliable execution
Agentic execution scales faster than verification or human adoption
AI output has expanded dramatically while enterprise adoption and verification tooling lag significantly behind. The asymmetry between cheap agentic execution and costly outcome validation is becoming the defining friction of the current deployment moment.
- Large org deployment requires navigating seven layers of internal process
- Execution cost falls fast while claim verification cost barely moves
- Recursive self-improvement tooling may signal actual exponential takeoff has begun
AI output is exploding while human adoption stays flat · @jenzhuscott Even working AI faces seven gates of operational hell in big orgs · Brandon Carl AI execution is getting cheap faster than verification can keep up · cgi.md Recursive self-improvement tools mark early exponential liftoff · Elad Gil The last six months may be among the most important in history · Elad Gil
One operator with parallel agents achieves team-scale output
A single person orchestrating 20 to 30 parallel agents can now match the throughput of engineering teams, investment analysts, and content operations. The emerging unit is one orchestrator setting objectives and reallocating compute, not a traditional team structure.
- One engineer shipped 40 pull requests a day using parallel agents
- In YC's spring batch, 60% of one-liners mention AI or agents
- The org chart flattens as compute replaces headcount scaling
Solo engineer ships 40 PRs a day running 20-30 parallel agents · Peter Yang (featuring Kun Cheng) A solo investor replaces a whole team with AI agent workflows · @zeroxkyle Naval's rule for the AI era is waste tokens to save human time · Naval Founders will run teams of agents like portfolio managers setting OKRs · deana Three-agent content pipeline cuts post production from 30 hours to 4 · Payton Kaleiwahea (summarizing Alex Lieberman) YC Spring '26 is the agents batch, 60% pitch AI in their one-liners · Ollie Forsyth Distill frontier models into local agents for a private personal assistant · Tomasz Tunguz
Open models match closed ones, pricing gap stays wide
The capability gap between open-weight and closed models has closed faster than expected while pricing has barely moved, creating immediate arbitrage for builders who route across providers. Application vendors that stay provider-neutral and charge on outcomes rather than tokens are finding structural cost advantages.
- Lindy's switch to DeepSeek cut costs and improved performance simultaneously
- Model routing becomes a primary lever for cost and risk management
- Charging on outcomes rather than inference tokens realigns incentive structures
Open and closed model capability converged, but pricing didn't · Chamath Palihapitiya AI apps beat the labs by staying Switzerland and routing across every model · Brad Menezes Lindy moves all traffic off Anthropic to DeepSeek v4, cheaper and better · Flo Crivello Raw intelligence will get cheap, but outcomes don't compress · Sierra AI conventional wisdom keeps reversing, from wrappers to desktop GUIs · Greg Isenberg OpenAI and Anthropic will move up the stack and absorb Cursor and Lovable · Jason Calacanis
Venture incentives misalign as fees rise and returns fall
Top VC funds now extract fees at many times their historical rates while returns have weakened relative to public markets. Early-stage culture has shifted toward performative input metrics, rewarding token burn and launch visibility over genuine product building.
- The venture power law at fund level is largely self-inflicted
- Benchmark's new growth fund signals a structural shift in firm strategy
- AI-native company speed is outpacing traditional slow-moving venture processes
Benchmark raises $2B and launches its first growth fund · Kate Clark AI investing hits cost walls as power-law belief breeds concentration · Brandon Carl Traditional VC's slow process clashes with AI-native company speed · Dave Fontenot Silicon Valley traded weirdos and tinkerers for input-maxxing strivers · Andrew Ziperski The venture power law is self-inflicted at the fund level · Dan Gray (Odin) Early-stage venture now rewards performative input-maximizing over building · Dave Fontenot Striver founders input-max performative metrics over real product fit · Andrew Ziperski Top VC firms now extract 10x more fees as their returns weaken · Dan Gray
Deterministic orchestration outperforms emergent multi-agent coordination
Structured, explicit workflows driving small agent loops are proving more reliable than non-deterministic multi-agent systems for complex tasks. New platform features and DSLs with primitives like parallel() and pipeline() are making this architecture the practical standard.
- Explicit workflow primitives replace emergent coordination between autonomous agents
- Structured orchestration unlocks multi-step tasks previously too complex for agents
- Context engineering, not model capability, is the remaining differentiator
Deterministic workflows orchestrating small agent loops beat agent soup every time · dex (@dexhorthy) New DSL gives subagents explicit workflow code instead of emergent coordination · Michael Livs (@micLivs) The /workflow feature cracks multi-step tasks agents long failed at · Jason Zhou Deterministic workflows around small agent loops beat multi-agent soup · dex Claude Code auto-generates orchestration plans it follows across hundreds of agents · cat
Infrastructure, not model quality, is now the real constraint
Model capability has crossed a threshold where harness, connectors, and reliable uptime matter more than raw intelligence. With nonhuman traffic exceeding half of some APIs, the build question has shifted from 'can the model do this' to 'can the system stay up and integrate.'
- No unified control plane yet exists for multi-domain agent orchestration
- Products lacking headless interfaces risk repeating the early mobile-era mistake
- Agents are beginning to manage their own threads, worktrees, and orchestration
Agents now need only a harness, connectors and reliable uptime, not better models · @nickbaumann_ No single control plane exists to manage agents across code, design and ops · Dennison Bertram Codex now manages its own threads and spins up worktrees for parallel tasks · Guinness Chen Agents are the next mobile, with 53% of one API's traffic already nonhuman · Gokul Rajaram Shopify's River is an AI agent living inside company Slack · shopify.engineering Sepo agent runs in GitHub Actions to implement issues and review PRs · github.com
Professional networks are becoming AI-queryable deal infrastructure
A set of tools now treats relationship graphs as structured, searchable data for finding warm backchannels, sourcing founders, and activating sales. The workflow collapses what previously required hours of manual research into a single natural-language query.
- Warm intros are closing more stuck deals than cold outreach
- Early founder signals surface from follower graphs before pitches arrive
- Network search now reaches across teammates' and friends' connections
Agents complement humans rather than replace them near-term
Spending on agents shows sharply diminishing returns beyond a baseline, while the competitive split is between people using AI and those who are not. The optimism and judgment required to act under uncertainty remain human advantages.
- Returns to agent expenditure fall off faster than for human labor
- The competitive divide is humans with AI versus humans without
- Intense conviction about outcomes remains a human edge agents cannot replicate
AI amplifies humans while expanding total work
AI tools compress months of elite analysis into hours and multiply individual leverage, yet paradoxically generate new layers of judgment-dependent work. As autonomous agents approach job-execution capability, protecting human attention emerges as a design priority.
- Autonomous agents may execute full jobs with screen memory and local models alone
- Exceptional individuals gain disproportionate leverage as AI commoditizes routine skill
- Design is bifurcating between tools that capture attention and tools that protect it
AI automation creates more work for humans, not less · every.to An agent that watches you work for a week then does your job is shipping now · aaronjmars AI makes exceptional people more important by amplifying human judgment · unknown AI compresses months of Citadel-grade analysis into hours · @Bobbxu A new tool category is being built to protect attention, not capture it · @scottbelsky
Capital concentration strains venture's exit pipeline
Capital is concentrating in large firms expanding into seed while fewer specialists fund consumer startups. The structural question is whether concentration produces enough exits and distributions to sustain the ecosystem.
- Consumer seed has few active writers in 2026
- Largest firms expand seed activity without widening the real opportunity set
- Early forced exits may become the norm rather than long-term compounding
Non-consensus bets capture outsized returns in finance
The biggest returns in venture and crypto both trace to non-consensus positioning. In venture that means being an information bottleneck on unpopular facts, while in crypto the largest outcomes come from new financial primitives and genuinely novel concepts.
- Technical papers and old books are the least-competed signal sources
- Incremental crypto improvements face customer acquisition headwinds that erode returns
Agents compose, persist memory, and self-evaluate through traces
The dominant architecture pairs a general orchestrator with specialists called as tool calls, while agents that write to memory after each task build compounding advantages. Reasoning traces, not raw compute, are the primary bottleneck for performant agentic systems.
- Forcing agents to render UI exposes errors they otherwise skip
- Traces enable automatic eval generation and self-debugging by other agents
- Running 100 continuous cloud agents signals post-scarcity developer tooling
Specialized experts like Harvey become tools the general agent pays to call · Nathan Baschez (@nbaschez) Agents that write to their own memory after every task compound an advantage · @gregisenberg OpenClaw runs 100 cloud agents reviewing every PR as if tokens are free · @steipete Reasoning traces, not compute, are the real bottleneck in agentic systems · Saneel (@sanlsrni) Agent traces let other agents auto-generate evals and debug themselves · ben hylak (@benhylak) Forcing agents to render a UI surfaces mistakes they would otherwise miss · @goodalexander
Agent-native tooling layer is forming and monetizing fast
Infrastructure designed specifically for agent consumption is taking shape: pay-per-use API catalogs, version control for agentic workflows, and multi-agent CLIs. Agents are becoming first-class economic actors with tooling needs distinct from human-facing software.
- Premium subscriptions can substitute for costly API calls in agent pipelines
- Version control for agents reached $35M market cap in two months
Hermes agents now search X via Premium instead of burning paid API calls · Kevin Simback (@KSimback) Git for agents Gitlawb hits $35M market cap in two months · @igoryuzo Grok Build runs a clickable CLI juggling many agents at once · Jason Ginsberg (@JasonBud) Ampersend launches a pay-per-use API marketplace built for AI agents · ampersend (@ampersend_ai) Claude debuts agent view, a unified dashboard for all coding sessions · Claude (@claudeai)
Venture fund sizing math is fracturing under round inflation
Series A inflation toward $30-100M forces dedicated seed funds to $300M+ or retreat to pre-seed, while megafund exit requirements have warped 'venture-backable' into a near-meaningless label. A $5B exit generates excellent returns for most funds but fails the pools that must clear $50B.
- Secondaries culture pushes founders toward tradable value over durable companies
- Anthropic voiding named secondary transfers signals active corporate counter-pressure
- Sub-$5M new-category bets outperform large raises in crowded proven markets
If Series A hits $30-100M, seed funds need $300M+ to stay relevant · @nunzi46 Selling-to-secondaries advice pushes founders toward tradable over durable · @arian_ghashghai Sub-$5M new-category bets beat $100M raises in crowded markets · Neil (@neilhar) Altos organizes funds around companies, heretical in venture · Ho Nam (@honam) Anthropic declares all secondary transfers of its stock illegal and void · Gabriel Shapiro (@lex_node) Venture-backable is meaningless when only megafunds need $50B exits · Arian Ghashghai (@arian_ghashghai)
AI dissolves execution moats, elevates taste and trust
AI has eliminated traditional product development bottlenecks, collapsing the execution gap that once separated high-performers. What remains scarce is taste, strategic judgment, and the trust built through genuine human presence in networks.
- VC analyst work is replaceable by agents; curated human networks are not
- Whether data or integration moats hold depends entirely on team execution
- The outlier founder rarely works on the consensus hot theme of the year
The hot theme rarely produces the year's most valuable company · Hadley Harris (@Hadley) Permanence is the rarest asset class in an age of AI disruption · Saneel (@sanlsrni) AI eats VC analyst work, leaving trust and network to humans · Rebecca Kaden (@rebeccakaden) AI startup moats range from real to false depending entirely on the team · Deedy (@deedydas) AI deleted the execution moat and pushed the constraint up to taste and judgment · Scott Belsky (@scottbelsky)
New primitives create market categories impossible before
Enabling technologies, from GPS to cryptographic receipts to multi-agent AI, each open a window to build categories that were structurally impossible before. Enterprise AI products and receipt-native economic instruments are two current examples of this dynamic playing out simultaneously.
- Proof of action as currency creates economic layers beyond existing financial rails.
- Supervisor-subagent-judge architecture is becoming the standard pattern for enterprise AI products.
- Early movers in platform windows capture disproportionate market position.
AI amplifies output and demand rather than replacing roles
Efficiency gains from AI increase rather than reduce demand for augmented roles, consistent with historical patterns of tool-driven expansion. The same dynamic applies individually, where AI leverage raises what a single person can produce in a given timeframe.
- Software engineering demand is rising alongside, not despite, AI coding tools.
- AI leverage can compress significant brand-building output into minimal weekly time.
- Displacement narratives miss the historical pattern of efficiency-driven demand expansion.
Agentic infrastructure is becoming the primary value layer
Applications are thinning as the infrastructure managing state, memory, context, and coordination becomes the load-bearing substrate of software. The harness, filesystem, and identity primitives around agents are beginning to matter more than the models themselves.
- Harnesses should evolve like org charts, not be architected once and built upon
- Memory is converging on git-backed files and grep, not proprietary knowledge graphs
- Agent identity may be the generational primitive this computing era has been missing
AI work shifts to a continuous agent runtime that humans and agents share · @danshipper When agents generate workflows the app goes thin and the database becomes runtime · @anshublog Autonomous development splits into reactive, exploratory, and user-spec modes · Peter Zakin The agent harness is an org chart you keep adjusting, not backend built once · Yohei Nakajima AutoGen and agent marketplaces are dead, context engineering compounds · Ronin Agentic infra makes people coalesce into a new kind of organization · Matt Huang Mesa gives agents a durable POSIX filesystem with branches, diffs, and rollback · Oliver Agent Identity is the new cell, ushering in Sovereign Computing · Josh Crumb Future computing splits into ambient intent, agentic execution, and ephemeral review · signüll AI memory converges on git files and grep, knowledge graphs losing · andrew pignanelli
Crypto becomes default plumbing for agent transactions
The dollar cannot subdivide below one cent and requires permissions across jurisdictions, making crypto the structurally necessary settlement layer for autonomous agent transactions. Stripe's recent crypto integration positions this as invisible infrastructure delivered under a trusted consumer brand.
- Buybacks may be the only credible path to token value accrual beyond speculation
- Platforms that own the customer relationship can now recapture interchange via stablecoins
Buybacks are the only credible path to token value accrual · Jeff Dorman Codex adds messaging and payments, pushing coding tools into financial rails · Brian Flynn Stripe ships crypto as agent-native plumbing under a trusted brand · Dan Romero (@dwr) Platforms reclaim bank revenue by issuing their own stablecoins, cards and FX · Lightspark Universal Basic Equity in crypto L1s replaces UBI in the AGI economy · Raoul Pal
AI eats incumbent moats faster than SaaS did
AI compresses defensibility timelines because the strongest traditional moats are absent from most AI products. Token-reselling and apathy-dependent business models face the steepest structural exposure.
- Cool interfaces and system prompts are far weaker moats than SaaS incumbents had
- Fintech profits built on customer inertia are structurally vulnerable to autonomous agents
- Law AI vendors built on token resale face disintermediation as labs go direct
Harvey and Legora are token resellers exposed if labs go direct · @willchen500 AI-mageddon hits harder than SaaS because the moats are weaker · @scottbelsky AI splits firms into hyper-optimized giants and nanocorps · Katie Chiou Fintech profits run on customer apathy, and AI agents will route around all of it · Anish Acharya
Pathological determination beats brilliance at founder selection
Pathological determination, present in under 1% of founders across thousands of evaluations, is the rarest and most predictive success trait. Most investors systematically overweight raw talent while underweighting obsession, which is anti-fragile where brilliance is not.
- Relentlessness without conviction just maintains someone else's order
- Teams weigh as heavily as product ideas in early-stage evaluation
- Box-checking VC frameworks would screen out many eventually category-defining founders
Across 5000 founder meetings, relentless determination was the rarest success trait · @adamshuaib Pathological determination beats brilliance and shows up in under 1% of founders · Adam Shuaib World-changers pair relentlessness with conviction, and either alone is a curse · Joel John Half of Speedrun's startup evaluation weighs the team, not the product · Jon Lai Running out of money rarely kills startups, it just hides the real cause · arian ghashghai Venture's box-checking scale game would have rejected a young Bezos or Musk · Dan Gray
Owning model and product together is the new moat
In AI-native markets, margin lives in the customer relationship rather than the code, so open-source becomes viable distribution while value stays upstream. Winning in AI coding requires closing the loop between model and product, as the Cursor-SpaceX deal signals.
- Open source is now wholesale pricing, not competitive disadvantage
- Markets repricing from P/E toward free cash flow as AI erodes terminal value
Agents are the second user of every interface
Infrastructure for agent-driven desktop automation, browser agents that learn through self-play, and pixel-streamed UIs with no DOM are converging on the same insight. Products now need to be designed for machine traversal as much as human use.
- Designing for agents means designing for workflow, not screens
- No-DOM pixel streaming eliminates the layout engine entirely
- Open-source desktop drivers now enable agent multi-player and multi-cursor
Cua open-sources a macOS driver letting any agent drive any app in the background · Cua (@trycua) Products now have two users, humans and the agents reading the interface · Simon Taylor (@sytaylor) Agents learn browser automation through self-play like AlphaGo · @shreypandya Flipbook streams every UI pixel from a model, no DOM or code · @zan2434
AI compresses knowledge work from hours to minutes
Content pipelines, essay workflows, and self-improving automation are cutting hours-long tasks to minutes. The competitive edge shifts from execution speed to conceiving workflows that non-practitioners cannot imagine.
- Giving models success criteria outperforms prescribing steps
- Understanding LLM internals enables conceiving entirely different product possibilities
- Multi-platform scraping can cut daily content creation by ninety percent
Four-step AI essay workflow: transcripts, interview, outline, draft · brett goldstein Content engine scrapes 9 platforms, cutting daily creation from 2 hours to 10 minutes · Ronin Karpathy drops free 3-hour course on how LLMs actually work · Rahul Karpathy-inspired claude.md hits 78.5K stars curbing silent assumptions and bloat · Om Patel (@om_patel5)
Venture capital broken across structure, culture, and returns
Retail VC funds stack three tiers of fees before any return reaches investors, while companies raising over $3B post negative IPO alpha. Terminology disputes and trauma-signaling in pitches signal a profession struggling to define itself.
- Seasoned founders build 'never work with' lists, not wish lists
- YC's moat is founder attraction, not application filtering
- Trauma signaling is replacing origin stories in founder pitches
VC argues endlessly because nobody agrees what its core terms mean · Dan Gray Founders now lead pitches with personal trauma instead of origin stories · Adam Draper (@AdamDraper) YC's real moat is attracting the next Coinbase, not filtering startups · Pareen (@pareen) First-time founders list dream funds, veterans list ones to avoid · pj (@BeingPractical) Retail VC funds charge hedge fund fees for VC illiquidity with mutual fund marketing · @sterlingcrispin Startups raising over $3B post negative IPO alpha of -63% · @credistick AngelList's 1% retail VC fund actually carries a 3.61% expense ratio · @covered_call
Autonomous agents operating without human oversight
Reasoning models now run hedge funds and sales pipelines with no human approval loops, while multi-model orchestration layers route tasks across dozens of models automatically. Security tooling, sandboxing, and firewalls mark the new threshold for production trust.
- Agents outperform domain experts on optimization without domain knowledge
- A config file and folder structure already constitute a working agent
- MCP servers are bridging AI agents into legacy business workflows
Perplexity launches a 24/7 AI operating system routing tasks across 19 models · Evan Luthra A reasoning model runs the hedge fund with no human managers · Michael Royzen (@MichaelRoyzen) Aeon ships a fully autonomous agent framework with zero approval loops · aeon (@aeonframework) OpenClaw is now production-ready with sandboxing, firewalls, and supervision · Garry Tan (@garrytan) A folder with a CLAUDE.md file already functions as an agent · @kieranklaassen On optimization problems, agents ignoring domain knowledge crush the experts · @danrobinson A LinkedIn MCP server lets AI agents run prospecting and outbound sales · denzell
Ambient AI compounds personal context over time
AI is shifting from reactive assistant to always-on layer that watches screens, listens to conversations, and accumulates personal context until it can surface a user's own past thinking. The value proposition is compounding over months, not single-session utility.
- Sitting on top of existing data (CRM, email, calendar) removes cold-start friction
- Replacing the static phone home screen with an ambient agent layer is shipping
Omi for Desktop is a local AI that watches your screen and advises next steps · kodjima33 Skye replaces the static iPhone home grid with an ambient agent layer · @signulll Claude plus Obsidian compounds context until it surfaces your own past thoughts · @yb_effect Portfolio monitor auto-calculates runway and sentiment from existing CRM and email · @yoheinakajima
AI deployment pace outrunning quality and incentive alignment
Model providers are quietly degrading reasoning quality, vibe-coded systems lack depth, and the ecosystem lacks go-to-market structures to align incentives around meaningful outcomes. Open source transparency is also becoming a liability as AI automates exploit discovery at scale.
- Quietly cutting reasoning effort without disclosure erodes ecosystem trust
- Transparency becomes exploitable when vulnerability discovery can be automated
- The real stakes may be human meaning, not job displacement
Vibe coding ships fast but great systems still take real time · Nikunj Kothari (@nikunj) Cal.com closes its core code as AI turns transparency into exposure at scale · Bailey Pumfleet (@pumfleet) AI lacks go-to-market and incentives, risking a generational fumble · @chamath Anthropic quietly cut Claude's default reasoning effort, telemetry caught the drop · Teng Yan AI's real challenge is human meaning, not job displacement · sam lessin