NARRATED · CAPTIONED · PRECISE ENGINEER
Narrated visual briefing
Find a workload-shaped wedge
Audio-only version
Search mode
Before building a platform, switch from execution mode to search mode. Steve Blank describes a startup as a search for a repeatable business model, and Y Combinator compresses the work into launching, talking to users, and iterating. For an inference company, your first launch may be a benchmark, a manual deployment, or a capacity rescue. The point is not self-service polish. The point is to expose a claim to a real operator and learn whether the failure matters enough to change behavior.
Interview protocol
Ask for the last concrete event. Walk me through the last model you shipped. Where did the schedule slip? What did the failure consume? What workaround did you try? A hypothetical question such as, would you pay for lower latency, invites politeness. A historical question reveals budgets, systems, decision makers, and urgency. End with a commitment test: if we remove this failure in four weeks, will you provide traffic, an owner, and a paid pilot?
Wedge score
Score candidate workloads, not broad industries. A useful wedge has a frequent painful event, visible existing spend, reachable buyers, representative data, and a natural expansion path. Real-time voice, private multimodal inference, long-context coding agents, and fleets of fine-tuned adapters are examples of workload shapes, not recommendations. Your founder access matters as much as technical elegance. A smaller problem with ten reachable operators is more valuable than a giant category you can only describe from the outside.
Evidence example
Imagine twelve teams tell you that faster inference sounds interesting. That is not twelve signals. Now imagine five coding-agent teams independently report the same event: long shared repository context is reprocessed after routing moves a session, time to first token misses the interactive budget, and engineers manually pin traffic to instances. Two teams share sanitized traces showing the repeated prefixes. One agrees to a four-week paid pilot if you improve p95 time to first token without changing answer quality. That is a real wedge candidate. The incident identifies a user, a workload shape, a broken control decision, a reproducible input, and a commercial gate. Your interview ledger should preserve the difference. Record compliments in a separate column and never count them toward validation. Count observed incidents, access to artifacts, named owners, and commitments. If the five incidents differ, do not average them into an artificial platform. Pick the cluster with the clearest consequence and strongest access, or keep searching.
Evidence gate
Use a hard gate before building the control plane. Five independent buyers should describe the same painful event. At least two should share a representative workload or traffic trace. One should commit to a paid or tightly time-bounded design partnership. This gate does not prove a company exists, but it turns architecture from imagination into response. Your assignment is a two-week, twenty-interview sprint and a ranked wedge memo based only on observed behavior.
WORKING MODEL
Four ideas to carry
Search, then execute
A startup is initially a search for a repeatable model. Your interview plan is an experiment plan, not a sales script.
Past behavior beats preference
Ask about the last outage, migration, bill, or launch. Hypothetical willingness to pay is weak evidence.
Workload beats vertical
“Healthcare” is broad. “Streaming pathology segmentation in a hospital VPC with a 200 ms p95” is a tractable product boundary.
A wedge has an expansion path
Start narrow enough to win, but pick a control point that naturally expands into routing, optimization, compliance, or capacity.
REFERENCE
Customer discovery for infrastructure
Infrastructure buyers rarely say “I need a new platform.” They report a missed launch, a cost shock, an SLO violation, a security review, or an engineer who spends every Monday nursing deployments. Interview around the event.
- Context. “Walk me through the last model you put in production.”
- Failure. “Where did the schedule slip or the system become unpredictable?”
- Cost. “What did the failure consume—engineer weeks, GPU idle time, lost requests, or delayed revenue?”
- Workaround. “What did you try, and why is it still unsatisfactory?”
- Commitment. “If we remove this failure in four weeks, will you provide traffic, an owner, and a paid pilot?”
YC's core loop is launch, talk to users, and iterate; Steve Blank frames this stage as searching through experiments rather than executing a fixed plan.E11–E12 For deep infrastructure, “launch” can be a benchmark report or a concierge deployment before it is a self-serve product.
Candidate wedges worth testing
Long-context coding agents
Repeated prefixes, bursty concurrency, high inter-token sensitivity. Potential edge: KV-aware routing and cache economics.
Private multimodal inference
Large payloads, custom preprocessing, VPC constraints. Potential edge: data plane installed in customer cloud.
Real-time voice
Strict latency budgets and streaming. Potential edge: admission control and regional warm pools.
Fine-tuned model fleets
Many LoRAs, sparse traffic, rollout risk. Potential edge: rapid adapter loading and safe promotions.
Regulated batch inference
Auditability and predictable completion windows. Potential edge: cost-aware scheduling and evidence logs.
Do not pick from this list because it sounds fashionable. Score only workloads where you can reach users, access representative traffic, and measure a failure better than incumbents.
The five-customer gate
Proceed to the technical MVP when five independent buyers describe the same painful event, at least two share a representative workload, and one signs a paid or tightly time-bounded design partnership. Until then, your job is learning.
SHIP EVIDENCE
Run the 20-interview sprint
Use the interview guide and wedge scorecard. Interview 20 operators in two weeks, summarize only observed events, rank three workloads, and ask the top five for traffic traces or replayable requests.
Course artifact: labs/customer-interview-script.md
DIAGNOSTIC CHECKS
Can you use the model?
CHECK 1 · application
Which interview question produces the strongest evidence?
CHECK 2 · discrimination
Which workload is best defined?
CHECK 3 · transfer
What is sufficient evidence to start a general platform build?
EVIDENCE LEDGER
Sources and limits
Vendor performance and product claims are labeled as first-party. Prices and product surfaces can change; follow the live links before making a purchasing or fundraising decision.
- ev-baseten-product Baseten overview · Baseten
Baseten presents a managed path from an open, fine-tuned, or custom model to a production API with containerization, GPU scheduling, autoscaling, observability, model-specific runtime optimization, and multi-cloud placement. - ev-yc-launch YC's essential startup advice · Y Combinator
Y Combinator advises founders to launch a useful early product, talk directly to users, iterate from observed needs, keep the team small before product-market fit, and care about unit economics rather than scaling an unprofitable product. - ev-blank-search Search versus execution · Steve Blank
Steve Blank distinguishes a startup's search for a repeatable business model from the later execution of a known model. Customer development turns business-model assumptions into experiments that can be falsified and revised. - ev-fireworks-serverless Serverless overview · Fireworks AI
Fireworks documents pay-per-token serverless inference with no GPU sizing and features such as prompt caching and session affinity. It is evidence that the commodity API layer already has mature specialist competitors. - ev-modal-platform Modal introduction · Modal
Modal offers general serverless compute with per-second billing, GPU functions, sandboxes, and elastic scale. It broadens the competitive set beyond companies that market themselves only as inference platforms. - ev-aws-provisioned Amazon Bedrock Provisioned Throughput · AWS
Amazon Bedrock lets customers buy fixed-cost model throughput in model units and commitment periods. Hyperscalers therefore compete for predictable enterprise inference capacity in addition to on-demand API traffic.