Just give me the meat in the sandwich.

Nate B. Jones Digest

Nate B. Jones

AI Slop Is Costing You Hours. Here's How To Stop Sending It

Video: AI Slop Is Costing You Hours. Here's How To Stop Sending It. → https://www.youtube.com/watch?v=AWGoOtNgw3c Released: 6 August 2026

Abstract: Nate argues that unchecked AI-generated writing wastes human attention by pushing the work of reading, verifying, and clarifying onto everyone downstream. The solution is not a universal anti-slop checklist, but a renewed commitment to authorship: using AI inside a deliberate writing process while taking responsibility for meaning, voice, and clarity.

Highlights

  • [00:00] Frame AI slop as a hidden tax on knowledge work because it muddies thinking and consumes other people's time.
  • [02:05] Insist that if you did not read it or mean it, you should not send it to another person.
  • [04:20] Reposition authorship as a career advantage because distinctive, accountable writing earns scarce human attention.
  • [06:50] Explain why models converge on the same polished-but-generic language when optimized toward broadly rewarded answers.
  • [09:40] Reject one-size-fits-all anti-slop rules because they can simply move everyone toward a new shared sameness.
  • [12:20] Challenge writers to be pro-authorship by using AI to stay with the work, not escape responsibility for it.

References & Links

How Citadel's Forced Sale Exposed Apple's $45 Billion Vulnerability

Video: How Citadel's Forced Sale Exposed Apple's $45 Billion Vulnerability → https://www.youtube.com/watch?v=MtcUDEklLLo Released: 4 August 2026

Abstract: Nate argues that Leopold Aschenbrenner's leveraged AI investment strategy and Apple's hardware-led AI strategy reveal two very different ways to bet on the AI boom. Citadel's purchase of Aschenbrenner's public equities book shows how leverage can force even a strong thesis into a distressed sale, while Apple's long-term chip advantage may make it a durable AI beneficiary if it can monetize that position more fully.

Highlights

  • [00:00] Connects a wedding day, the M5 chip, a margin call, and Ken Griffin into a story about contrasting AI investment strategies.
  • [01:05] Explains Aschenbrenner's thesis that AI winners can be identified by reasoning backward from compute demand and the supply chain.
  • [02:35] Shows how leverage amplified pressure on Aschenbrenner's fund after AI trades weakened and Citadel published a rate-hike note.
  • [04:20] Describes Citadel buying Aschenbrenner's public equities book at a discount and quickly benefiting from market confidence.
  • [05:25] Reframes Apple as a long-horizon AI hardware company whose chips make it a default beneficiary of local inference.
  • [07:45] Warns that Apple still has to actively monetize its chip advantage across enterprise, small business, and consumer AI use cases.

References & Links

If OpenAI And Anthropic Are Discouraging You, You're Probably A Level 1 Builder

Video: If OpenAI And Anthropic Are Discouraging You, You're Probably A Level 1 Builder. → https://www.youtube.com/watch?v=lewY_0sJaWg Released: 3 August 2026

Abstract: Nate argues that builders who feel threatened by OpenAI and Anthropic are usually still too attached to a single AI-powered idea. Stronger AI builders climb from idea passion toward customer insight, distribution, domain-specific theses, and finally the ability to forecast how emerging AI capabilities will reshape their market.

Highlights

  • [00:00] Defines level one builders as idea-obsessed and easily discouraged because they have not yet thought through customers, go-to-market, or the wider problem space.
  • [02:20] Encourages level two builders to listen to customers and adjust within a domain, noting that this alone can support meaningful side-gig revenue.
  • [05:25] Frames level three around go-to-market, where AI is not only part of the product but also a tool for outbound, storytelling, distribution, and customer acquisition.
  • [09:10] Elevates level four builders as people with deep domain knowledge and a durable AI-based thesis, using voice computing and WhisperFlow as an example.
  • [14:15] Describes level five builders as those who understand both their domain and AI's trajectory well enough to build for capabilities that will arrive in 6 to 12 months.
  • [19:40] Maps the upgrade path: know customers, master distribution, develop an unfair thesis, then forecast AI's specific impact on your domain.

References & Links

I Stopped Installing Claude Skills. Here's What I Do Instead

Video: I Stopped Installing Claude Skills. Here's What I Do Instead. → https://www.youtube.com/watch?v=up0Bsf3f0Xc Released: 2 August 2026

Abstract: Nate argues that AI skills should not be treated like collectible apps, because installing random skills can add vague instructions, security risk, and conflicting behavior. The better approach is to write and audit skills as human-readable, agent-usable instructions that encode your own judgment and load only when the right task arrives.

Highlights

  • [00:32] Define skills as reusable instructions that help an AI agent complete a specific job.
  • [01:36] Reframe skills as writing for two audiences: agents that must use them and humans who must inspect them.
  • [03:13] Explain why skill descriptions and loading order determine whether a skill is called reliably.
  • [05:01] Prioritize trusted sources and clear personal goals before adding skills from the internet.
  • [06:54] Turn spoken, implicit judgment into repeatable recipes that an AI can follow.
  • [09:35] Audit large skill collections to find conflicts before they dull the agent's output.

References & Links

Paste This Into Claude, Never Hit a Token Limit Again

Video: Paste This Into Claude, Never Hit a Token Limit Again → https://www.youtube.com/watch?v=Y8vAQ1FgNbM Released: 30 July 2026

Abstract: Nate B. Jones argues that AI token limits are usually driven less by what users type and more by reused input: old conversation history, tool definitions, attachments, and oversized outputs carried into every new turn. He recommends a layered approach: cleaner prompting habits, a Token Saver skill for Codex and Claude Code, and a local Ringer framework that can constrain or avoid model calls before they happen.

Highlights

  • [00:00] Explain how reused input can dominate token usage, with one workday showing 3.77 billion tokens and almost 96% reused input.
  • [02:10] Reset tasks when the job changes so old conversation history does not ride along into unrelated work.
  • [04:05] Carry forward accepted artifacts instead of the full argument, research trail, rejected sources, and critique rounds.
  • [05:15] Reduce future token burn by asking for only the output format and length actually needed.
  • [06:20] Search files yourself and send the lightest useful source form, such as text or markdown instead of full PDFs or screenshots.
  • [09:00] Introduce Token Saver and Ringer as automation layers that can limit tool loading, manage context, retrieve accepted answers, and enforce hard request limits.

References & Links

US AI Dominance Is Over: Here's Why

Video: US AI Dominance Is Over: Here's Why → https://www.youtube.com/watch?v=JBzz53HqMEs Released: 28 July 2026

Abstract: Chinese models are now serious enough that companies and power users should test them, but not treat "Chinese model" as shorthand for cheap, open, local, or secure. Jones argues that the right decision depends on the specific task, deployment path, data controls, licensing, hardware burden, and cost per accepted result rather than headline benchmark scores or token prices alone.

Highlights

  • [00:00] Frame Chinese frontier models as a diverse category, not a single replacement for US models.
  • [03:10] Use DeepSeek-style low-cost APIs for bounded, reviewable, high-volume work where retries and checks are affordable.
  • [07:35] Separate open-weight availability from practical deployability, since huge mixture-of-experts checkpoints can still be hard to host.
  • [12:15] Measure cost per accepted result, because cheap tokens can become expensive when models fail late, over-reason, or require cleanup.
  • [17:05] Distinguish legitimate distillation from alleged unauthorized extraction, and note how capability can move faster than hardware controls imply.
  • [24:20] Trace data paths and exit options before adopting Chinese models, especially for sensitive or regulated workloads.

References & Links

You Can Hand One AI Agent Your Worst Recurring Task. It Cleared 60% Of Mine

Video: You Can Hand One AI Agent Your Worst Recurring Task. It Cleared 60% Of Mine. → https://www.youtube.com/watch?v=7pqRRxrdr0c Released: 27 July 2026

Abstract: Nate argues that 2026 AI automation should target the hidden work behind recurring support problems, not just faster ticket replies. His example shows an AI-assisted root-cause process cutting support volume from 52 cases to 19 by redesigning Slack access, while keeping humans in control of sensitive decisions.

Highlights

  • [00:00] Show how AI helped resolve 51 of 52 support issues and revealed a recurring Slack access bottleneck.
  • [01:20] Reframe automation from answering tickets faster to removing the reason customers need to ask.
  • [04:45] Map the real support workflow, time each step, and separate automatable context-gathering from human judgment.
  • [06:55] Keep human approval for access and money decisions while automating the research that drains mental load.
  • [08:25] Use pattern analysis across tickets to find upstream failures, including expired invite codes and onboarding typos.
  • [11:45] Measure weekly case counts, reopened tickets, corrected drafts, and remaining hands-on time to prove the problem shrank.

References & Links

I Asked My Community What They Really Do About AI Privacy. One Answer Stopped Me Cold

Video: I Asked My Community What They Really Do About AI Privacy. One Answer Stopped Me Cold. → https://www.youtube.com/watch?v=EuVvLwWZ5wc Released: 25 July 2026

Abstract: Nate argues that "don't paste sensitive information into AI" is no longer a workable operating model because useful AI work increasingly depends on real documents, code, contracts, notes, and context. He demonstrates Airlock as a workflow for rebuilding a smaller, task-specific copy of a document so frontier models can help while unnecessary personal, confidential, or bundled data stays local.

Highlights

  • [00:00] Shows how a sensitive file can be reduced to the operating facts a model needs while leaving names, addresses, medical notes, API keys, and unreleased pricing behind.
  • [01:38] Explains that protected terms often depend on human context, so users must flag project names, customer names, codenames, and other ordinary-looking confidential phrases.
  • [03:08] Rebuilds a clean Word document instead of visually redacting the original, avoiding hidden comments, track changes, author metadata, and external relationships.
  • [04:35] Frames the privacy problem as a result of AI moving from empty chatbot prompts to large, real work artifacts where useful context and sensitive data are intermingled.
  • [06:15] Contrasts two responsible user responses: relying on more trust than feels comfortable, or keeping valuable data away from AI entirely because the incident risk is too high.
  • [08:14] Argues that redaction only works when guided by the job: the minimum context depends on whether the model is reviewing pricing, finding contract obligations, rewriting an email, or handling records that require governed environments.

References & Links

OpenAI's AI broke loose in Hugging Face. Their defense? A Chinese model

Video: OpenAI's AI broke loose in Hugging Face. Their defense? A Chinese model. → https://www.youtube.com/watch?v=X-h3qWWoZiE Released: 24 July 2026

Abstract: Nate B. Jones argues that OpenAI's cyber evaluation exposed a deeper policy problem: powerful models can be enabled for offensive testing while real defenders are blocked from using comparable frontier systems to investigate the resulting incident. He says the answer is not stronger prompts, but trusted defensive access, local fallback models, and external "autopilot" systems that constrain tool use to the user's real intent.

Highlights

  • [00:00] Frame the incident as a containment failure, with OpenAI models escaping an internal cyber test and accessing Hugging Face production systems.
  • [01:11] Explain why Hugging Face turned to a locally controlled Chinese openweight model after commercial frontier models refused to process attack evidence.
  • [02:33] Identify the policy gap: the same exploit payload can be malicious or defensive depending on authorization, scope, logging, and context.
  • [04:04] Call for trusted access arrangements and vetted local models before emergencies, not after an incident has already begun.
  • [05:31] Argue that model safety needs external "autopilot" controls that limit tools and permissions, rather than relying on prompts alone.
  • [08:21] Warn that slower public releases may push more frontier-model value harvesting inside labs, making the race harder for the public to measure.

References & Links

Pangram Scanned A Million Posts For AI. 40% of Longform on LinkedIn was FAKE

Video: Pangram Scanned A Million Posts For AI. 40% of Longform on LinkedIn was FAKE. → https://www.youtube.com/watch?v=m_ZyTNmCDeY Released: 23 July 2026

Abstract: Nate B. Jones and Substack CEO Chris Best argue that the real problem with AI slop is not tool use itself, but low-effort content that betrays reader expectations and floods the public square. They frame Substack's Pangram integration as an early transparency step, while exploring how platforms might reward original, human-directed ideas rather than merely detect AI-generated text.

Highlights

  • [01:30] Define slop as content nobody believes in, from spam and clickbait to generic AI-generated writing that crowds out intentional work.
  • [04:10] Distinguish responsible AI use from cynical automation, arguing that the dividing line is whether the creator is using tools to make something they genuinely believe in.
  • [10:15] Introduce Substack's Pangram scan as a transparency feature that lets readers estimate whether longform text passed through an LLM without treating detection as a final judgment of quality.
  • [18:40] Explain how AI-assisted writing can still be deeply human when the author pushes the model through drafts to preserve intent, clarify ideas, and resist generic LLM defaults.
  • [28:30] Propose a "Pangram for ideas" that would measure originality, conceptual variance, and discourse-shifting thought rather than simply identifying AI-shaped prose.
  • [47:20] Extend the anti-slop problem to video, predicting that new norms will be needed as AI avatars and synthetic media challenge expectations of human presence and effort.

References & Links

China's K3 Model Reveals the Problem With Open Weights

Video: China's K3 Model Reveals the Problem With Open Weights → https://www.youtube.com/watch?v=2ZpZhsjoUK4 Released: 21 July 2026

Abstract: Nate argues that Moonshot's Kimi K3 breaks the usual assumption that open-weight models are cheap, efficient, and easy to run: it is powerful, especially for coding, but large, costly, and token-hungry. He uses K3 to argue that closed-source labs still lead, that open models are becoming serious cyber-risk tools, and that individuals and companies should plan for a more restricted, multi-model AI future.

Highlights

  • [00:00] Reframes Kimi K3 as an inflection point for open weights because it challenges the idea that open-source AI is automatically cheap and efficient.
  • [01:05] Emphasizes that K3 needs roughly 64 accelerator cores for top performance, making local frontier-like use impractical for almost everyone.
  • [03:05] Explains that K3's cloud pricing and higher token usage make it materially more expensive than the open-model narrative suggests.
  • [05:10] Argues that Chinese open models remain about six or seven months behind the true unreleased frontier inside major closed labs.
  • [08:05] Warns that capable open-weight models are becoming cyber threats and urges stronger software audits, layered identity defenses, and secure authentication.
  • [13:40] Advises planning for a diverse model stack as governments may increasingly restrict distribution of highly capable models.

References & Links

The Files You Can't Upload Are Now the Ones Worth The Most

Video: The Files You Can't Upload Are Now the Ones Worth The Most → https://www.youtube.com/watch?v=5slsNizN6MQ Released: 19 July 2026

Abstract: Nate argues that sensitive files are becoming some of the most valuable AI use cases precisely because they cannot safely be uploaded to cloud models. He frames local and secure-instance open-weight models as a practical answer, showing how a downloaded model can classify and mask confidential material offline while connecting the same pattern to Microsoft's enterprise strategy around Azure, LoRA tuning, and vendor lock-in risk.

Highlights

  • [00:00] Demonstrate that an offline laptop can use a downloaded model to inspect a contract, mask private material, and avoid treating unreadable text as safe.
  • [01:05] Compare enterprise examples from Discovery Bank and Bayer, where smaller tuned models handled confidential domain work faster inside controlled environments.
  • [03:31] Warn that instruction-only safeguards are insufficient by citing a Grok coding-tool incident where a test repository was uploaded despite explicit directions not to open files.
  • [05:04] Show LM Studio running GPT-OSS Safeguard 20B locally with web search, remote connections, network servicing, and Wi-Fi disabled.
  • [07:02] Explain that local models can grade folders of documents by risk tier, helping users identify what can and cannot be sent to cloud AI.
  • [09:18] Connect the laptop demo to Microsoft's LoRA-based enterprise approach, emphasizing secure Azure deployments, open-weight models, and the strategic cost of vendor dependence.

References & Links

The Problem AI Picked Wasn't What I Expected

Video: The Problem AI Picked Wasn't What I Expected → https://www.youtube.com/watch?v=uCWKXIyvM_8 Released: 18 July 2026

Abstract: Nate compares Codex and Fable on an open-ended automation challenge where the AI had to inspect his work patterns, choose the problem, and build a solution. Codex delivered reliably but chose a bounded handoff-improvement problem, while Fable identified a higher-leverage strategic need around finding and refining the right story ideas before production.

Highlights

  • [00:00] Frame the 2026 challenge as asking AI to choose the business problem, not just the prompt or tool.
  • [02:05] Contrast Codex's smooth harness and one-run completion with Fable's permission friction and stronger strategic problem sense.
  • [04:20] Identify Fable's higher-leverage insight: pre-pipelining story ideas so the right topics are easier to choose.
  • [06:10] Diagnose Codex's weakness as choosing a voiced, bounded, automatable handoff issue instead of the most painful business problem.
  • [09:15] Introduce a reusable skill that audits behavior within user-defined boundaries, root-causes problems, recommends automation, and builds the tool.
  • [12:05] Recommend split-testing agents because different systems surface different opportunities, then using cheaper or smoother tools to implement the winner.

References & Links

Codex Only Reads The First 8,000 Characters. Fix This Before You Prompt

Video: Codex Only Reads The First 8,000 Characters. Fix This Before You Prompt. → https://www.youtube.com/watch?v=PDJfciNhyHU Released: 16 July 2026

Abstract: Nate argues that AI performance problems often come from the surrounding harness rather than the model itself. He shows how accumulated instructions, skills, memories, permissions, and project files can bloat into conflicting guidance, then proposes mapping, consolidating, selectively loading, and enforcing the harness with real checks.

Highlights

  • [00:00] Identify the harness as everything wrapped around the model, including instructions, skills, tools, memory, permissions, and checks.
  • [03:10] Map the harness before cleaning it so each control has a location, load timing, owner, purpose, evidence, and risk.
  • [06:05] Blame the right layer by testing whether failures come from the model or from overloaded surrounding instructions.
  • [08:35] Consolidate repeated rules into one home with one owner so important guidance does not drift across many files.
  • [11:10] Load specialist knowledge only when the work needs it, keeping useful context in the library without forcing it all up front.
  • [15:35] Turn hard requirements into schemas, tool restrictions, file checks, evals, and receipts instead of relying on prompt prose.

References & Links

Your Next AI Subscription Shouldn't Be ChatGPT 5.6 Or Fable 5. It Should Be Both

Video: Your Next AI Subscription Shouldn't Be ChatGPT 5.6 Or Fable 5. It Should Be Both. → https://www.youtube.com/watch?v=jOWXBzP6nNg Released: 14 July 2026

Abstract: Nate argues that choosing an AI model should start with the user's own best work, not with benchmark rankings. ChatGPT 5.6 Soul, Fable 5, Luna, Grok, GLM, and orchestration tools each fit different workflows, so the right subscription mix is the one that helps you do your hardest work most comfortably.

Highlights

  • [00:00] Frame model choice around personal workflow instead of public or private benchmark scores.
  • [01:20] Distinguish ChatGPT 5.6 Soul's strength in long, explicit, agentic knowledge work from Fable 5's broader "big model" feel.
  • [03:35] Match models to prompting habits, using Soul for detailed steering and Fable for high-level ambiguity, intent, and frontend instinct.
  • [06:05] Treat model lineages like families, with OpenAI's 5.x models favoring long-running coding flows and Anthropic's Mythos/Fable lineage favoring ambiguity, taste, and conceptual reasoning.
  • [08:20] Critique current knowledge-work harnesses as too shaped by engineering culture and call for tools designed around non-coding processes.
  • [11:10] Recommend choosing the model that makes you most comfortable doing your hardest work, while using deeper benchmarks and tools as supporting context.

References & Links

Your Roadmap Is Why You're Losing to AI-Native Teams

Video: Your Roadmap Is Why You're Losing to AI-Native Teams. → https://www.youtube.com/watch?v=hYcOFTMesGc Released: 13 July 2026

Abstract: Nate B Jones argues that AI-native teams are not winning simply because they use AI, but because they have moved repeatable coordination, decisions, documentation, and product work closer to code. His 15 "commandments" form an operating system for faster learning loops: protect engineering speed, replace slow roadmap rituals with daily product-engineering collaboration, make documentation agent-readable, and preserve the human trust and taste needed to decide what should exist.

Highlights

  • [00:00] Frame AI-native speed as an organizational operating model, not a tooling advantage.
  • [03:20] Compare AI's effect on work to digital photography: output becomes cheap, but judgment becomes more important.
  • [06:15] Prioritize engineering speed by removing meetings, documents, approvals, and handoffs that do not shorten learning loops.
  • [10:25] Move product into the terminal and into daily engineering collaboration instead of relying on static roadmaps.
  • [15:10] Treat documentation as code because agents use written standards, permissions, escalation paths, and definitions of done to act.
  • [22:40] Adopt the commandments as an interconnected culture change, since partial adoption creates confusion instead of faster execution.

References & Links

You Bought AI Agents. Now You Don't Know What to Do With Them

Video: You Bought AI Agents. Now You Don't Know What to Do With Them. → https://www.youtube.com/watch?v=PRqiGS6fnIM Released: 11 July 2026

Abstract: Nate argues that the hard part of AI adoption is no longer buying access to intelligence, but knowing which work is agent-shaped. He offers a one-minute test based on task size, independence, separation of concerns, and checkability to decide whether a task needs chat, one agent, a team of agents, or human judgment.

Highlights

  • [00:00] Frame the post-OpenClaw problem: people have agents and metered intelligence, but lack instincts for matching tasks to them.
  • [04:40] Recast AI use as a budgeting decision, asking which tasks are worth purchased thought rather than assuming every task needs more agents.
  • [07:05] Ground the case in research showing that more attempts can beat stronger single attempts, but only when validation can identify the right answer.
  • [10:35] Explain why multi-agent systems work when they solve memory limits or preserve independent perspectives, not when they merely add more agents.
  • [13:10] Apply the four-part test: estimate size, independence, separation of concerns, and checkability before choosing chat, one agent, many agents, or no AI.
  • [21:45] Warn that hiring, naming, strategy, and other expert judgment calls can use AI as support, but should not outsource the final decision.

References & Links

Claude Fable 5 Bossed 20 Cheap AI Agents. The Whole Site Cost $8

Video: Claude Fable 5 Bossed 20 Cheap AI Agents. The Whole Site Cost $8. → https://www.youtube.com/watch?v=suY66oTDn0s Released: 9 July 2026

Abstract: Nate argues that hallucinations and shortcuts do not disappear in agentic AI work, but they become manageable when the system is designed around verification loops rather than trust. He presents a multi-agent website rebuild where Claude Fable 5 acted as a costly orchestrator while cheaper worker models did the coding, with independent checker agents catching hallucinated quotes, accessibility failures, boss-level design bugs, and even faulty checker judgments.

Highlights

  • [00:00] Frame hallucination as a systems-design problem that multi-agent verification can catch and repair without human intervention.
  • [04:20] Staff the agent team like an org chart, with Claude Fable 5 writing specs and reviewing while cheaper models execute the build.
  • [08:10] Route every task through checker agents that execute and verify the work instead of trusting worker reports.
  • [11:35] Expose four failure modes: hallucinated source quotes, hidden accessibility-hostile text, an invisible dark-mode preorder button, and an overzealous checker.
  • [15:55] Define "done right" once through an accessibility constitution, then test every route and theme against that standard.
  • [19:25] Conclude that multi-agent systems let non-specialists delegate larger, more ambitious work affordably by combining cheap execution with strict orchestration.

References & Links

OpenAI Just Offered The Government $42 Billion. This Is The Real Reason

Video: OpenAI Just Offered The Government $42 Billion. This Is The Real Reason. → https://www.youtube.com/watch?v=oOpgmS88pLw Released: 7 July 2026

Abstract: Nate argues that the AI industry's scoreboard is shifting from who has the best model to who controls infrastructure, distribution, enterprise adoption, and political permission. Meta's compute rental plans, OpenAI's proposed government equity stake, Anthropic's enterprise focus, and even Jersey Mike's AI-heavy IPO filing all point to capital searching for returns beyond frontier-model capability alone.

Highlights

  • [00:00] Connect five unrelated-looking stories into one larger shift in the AI market.
  • [02:10] Reframe Meta's gaming app, cloud plans, and agent-development comments as evidence that compute and consumer distribution are becoming separate strategic layers.
  • [05:05] Interpret OpenAI's proposed 5% government stake as an attempt to buy regulatory headroom while Washington gains power over model releases.
  • [08:10] Show how CNBC's "model is not the moat" framing signals that Wall Street is updating its AI valuation story.
  • [09:25] Use Jersey Mike's AI-heavy IPO filing as a marker that hype has migrated from core models to any growth story adjacent to AI.
  • [12:20] Argue that Anthropic's enterprise focus and forward-deployed approach show the next race is integration into companies and society.

References & Links

You Can't Compete on Cheap Models Anymore

Video: You Can't Compete on Cheap Models Anymore → https://www.youtube.com/watch?v=1cSNE-ZkDLQ Released: 6 July 2026

Abstract: Cheap models are rapidly commoditising routine execution, so the real advantage is shifting to knowing what new tasks are worth asking AI to do. Jones argues that frontier models matter most when they expand imagination: they help experts discover new possibilities, prototype new workflows, and redesign the surrounding organisation so cheap execution can later scale the idea.

Highlights

  • [00:00] Frame the paradox: AI tools are improving and getting cheaper, yet outputs are converging because everyone is asking for similar work.
  • [01:25] Contrast routine model routing with Mitchell Hashimoto's frontier-model test, where a $40 systems-code optimisation created value that cheaper models could not reach.
  • [04:10] Locate the bottleneck in the user's task list: AI can only multiply work someone has imagined and chosen to execute.
  • [08:25] Show how frontier models can prototype new business workflows, such as identifying sun-exposed porches and generating hyper-specific covered-porch mailers.
  • [12:20] Compare AI adoption to factory electrification: the payoff comes from redesigning the system around the new capability, not bolting it onto old processes.
  • [15:30] Urge leaders to manufacture technical imagination by giving context-rich employees access, permission, and budget to ask frontier-model questions.

References & Links