The old internet covenant — Yahoo gives you a free stock quote if you give it your eyeballs for an E-Trade ad — has been broken by agents. In a world where human eyeballs are no longer looking at things, what marketing looks like is a very open question.
speaker's argument; the future-of-marketing question is explicitly open
Randy Hunt explains that some insights are light but important, and that a designer can now quickly get meaningful input in real time by querying research agents, pulling as much into their working process as they want — versus the old slow process of asking the research team in Slack.
confident
Randy Hunt says Notion has been able to dramatically reduce the amount of ineffective research — research done well from a practices standpoint but that struggles to impact product work — by making insights available via agents.
confident
Before defending your position in a review, identify the other person's concern out loud (labeling), then stop talking — the tension usually drops and real problem-solving begins. The narrator attributes the technique to Chris Voss and says it works as well on stakeholders as across a negotiating table, because nobody can hear your argument until they feel heard.
Sujee recommends employing routers that are cache aware: the router knows where the caches are stored and routes requests accordingly, giving good cache hits and very good speed.
Carmen Li explains asset valuation as discounting forward-looking cash flows to time zero: for example, a three-year lease on A100s at a given rate, discounted back, yields the asset's worth.
hypothetical framing
Tenex incentivizes AI strategists perfectly at odds with engineers: AI strategists are incentivized for customer retention and growth, which prevents engineers from sandbagging story points or shipping bad code.
A revenue trajectory that reads as a good pre-AI business (~$1M year one, $2.5-3M projected year two) can still fall short of venture scale post-AI — benchmark against the new bar, not the old one.
described
Don't pivot mid-cycle in a seasonal business — you can't pivot after April 16th once you've committed to doing customers' tax returns, or in January after signing half your partners.
discouraged
Run a monthly automated wholesaler price-change analysis for a restaurant: compare current wholesaler prices against prior prices to catch 20-30% price increases and shift orders to cheaper wholesalers.
endorsed
In a cyclical/seasonal business, schedule the pivot decision inside the narrow window when it's possible — for their tax company, only about a month after October 15th allowed living in both universes; after that, the door closes.
described
Carmen Li explains that token forward contracts are like dedicated deployment or on-premise deployment: you reserve a bare metal GPU cluster with dedicated token deployment on top, getting guaranteed SLA for a particular model. With token forward contracts, quantization and precision are fixed for a period of time, solving the on-demand token API volatility pricing problem.
This is the creator's explanation of her own product's mechanics.
Carmen Li says Compute Exchange verifies GPUs, unlike other so-called brokers. For existing reserve capacity, verification includes checking UIDs, inventories, and performances, running NVIDIA DGX, InferenceX, and MLperf benchmarks. For forward capacity, they look through purchase orders and permits to ensure delivery at the desired level. They also do KYB and KYC verification of clients because new clouds worry about whether clients can pay at a scheduled rate.
Descriptions of internal company processes, as stated by the founder.
Use CamelCamelCamel, a free Amazon price-tracking website, to get emailed when a product drops to your target price — then complete the purchase yourself. CamelCamelCamel has 20 million American users and has operated as an affiliate website for 18 years — evidence that agent-like price-watching behavior already exists at scale. The best way to predict the future is to observe the present — look for existing behaviors at scale (like CamelCamelCamel) as signals of where agent technology is heading.
user count and age figures flagged as possible numeric errors
The old internet covenant — Yahoo gives you a free stock quote if you give it your eyeballs (via ads) — has been broken by agents. In the described example, Yahoo pays for a stock quote feed from NASDAQ, the user gets a free stock quote, and Yahoo gets the user's eyeballs on an E-Trade ad. People whose businesses depended on the eyeball covenant are adversarial toward agents, but there's a big opportunity to be synergistic — 'it's like the new SEO.'
speaker's interpretation
You need audit logs, a posture that keeps privilege information from being shared with the agent, and cost controls for the CFO.
endorsed requirements
Enterprise adoption requires audit logs, a posture that allows privilege information not to be shared with the agent, and cost control for the CFO so they understand where money is being spent and that tokens are being used efficiently. Enterprise adoption of AI agents will take more human time to adapt because you're still selling into a Fortune 100 company that has to be run with a certain level of diligence. As you move down market toward SMBs and mid-market firms, there are fewer security and compliance requirements and adoption will be much faster.
The go-to-market journey starts by getting a few people at any company to connect just their email and calendar — not core systems of record — to feel enough magic. After demonstrating value, you ask whether you've earned the right to ask for CRM connection and other tool integrations.
endorsed approach of the speaker's company
If you were Amazon, don't block AI agents — instead build a separate, excellent ad-supported browsing experience that's worth visiting on its own, and launch your own personal agent (a 'better Muse') with an exclusive on shopping, bundled into Prime.
Randy Hunt says his strongest factor when solving design problems is prioritizing understanding the people of the place — their behaviors and norms — as the most impactful context. He recommends paying attention through deep ethnographic research, longitudinal studies, or simply walking around the world with your eyes open, seeing the world as it is rather than as you want it to be.
No published protocol matches this selection.
The creators behind most of what cleared.
195 cleared insights match
99 cleared insights match
85 cleared insights match
Choose sources → to get a library of your own and a daily brief.
