Recommendation · MoneyShaun Crimmins recommends mapping each value touch throughout the pilot in advance, planning exactly when you deliver value and what your ask will be. In week one, improve a metric the customer didn't know they had (e.g., knowledge base coverage from 30% to 90%) and hand them the improved asset at the end of the pilot regardless of outcome. Use the improved metric from week one as leverage to ask for access to additional tools or integrations in week two.
30 Minutes to President's Club | No-Nonsense Sales · #607 - How to Turn Every Sales Pilot Into a Closed Deal | Shaun Crimmins
Recommendation · MoneyShaun Crimmins recommends including three slides before dropping into a demo: the biggest problems learned in discovery, how you solve those problems, and what success looks like today. Set up your demo so it could work without you in the room.
30 Minutes to President's Club | No-Nonsense Sales · #607 - How to Turn Every Sales Pilot Into a Closed Deal | Shaun Crimmins
Recommendation · MoneyBefore launching an AI agent, ensure it works exceptionally well, not just 'good enough', to avoid high churn. Many AI agents are average or require babysitting and manual work, leading to high churn rates of 20-40% per month. If you overpromise on an AI agent's capabilities, customers will cancel within a month or two. Do not build an AI agent that is solidly mediocre or only does 80-90% of the job; it's not enough to replace a human. Tune and optimize the AI to be exceptional. Do not ignore the cost of the AI agent; it could hit your profit.
Startups For the Rest of Us · Episode 847 | What Second Time Founders Do Differently, Pricing AI Agents, and More Listener Questions (Rob Solo)
Recommendation · MoneyRob Walling provides techniques to overcome fear of social interaction: hire a therapist or coach to join calls initially, use an LLM to draft outreach emails and prepare call scripts, practice sales conversations out loud (record yourself, use a mirror, practice while driving), and take small progressive steps. He emphasizes that repeated practice trains the nervous system not to go into fight/flight/freeze mode, and reminds that social interaction is not a life-or-death threat.
Startups For the Rest of Us · Episode 849 | Legal Threats, AI as Platform Risk, Building Confidence, and More Listener Questions (Rob Solo)
Recommendation · MoneyDerrick Reimer recommends using an LLM (e.g., Claude Code) to help a new developer understand the codebase by asking questions about major subsystems and reading Git history. He also suggests pair programming for 1-2 hours a day for the first couple of weeks to infuse team ethos and build rapport.
confident
Startups For the Rest of Us · Episode 850 | How Not to Screw Up When Hiring Your first Developer and More Listener Questions (with Derrick Reimer)
Recommendation · MoneyErik Allebest advises not assuming a growth spike from a macro event (like COVID or The Queen's Gambit) is permanent; wait to see if the baseline holds. He describes the 'white-knuckle' scalability issues across product and tech during rapid growth, and notes that growth became more predictable after building the right team.
No Priors: Artificial Intelligence | Technology | Startups · What Chess.com Teaches US About Superhuman Capabilities, with CEO Erik Allebest
RecommendationHodak recommends pairing a long-term research project (e.g., bio-hybrid neural interfaces) with a near-term commercial product (e.g., retinal prosthesis) to sustain the business, and emphasizes that having a real business is critical for a moonshot.
No Priors: Artificial Intelligence | Technology | Startups · From Restoring Sight to Reimagining the Brain, with Max Hodak
Recommendation · BodyChoose exercises where the arm finishes back behind the body, putting the long head in its most shortened position.
ATHLEAN-X™ · STOP, You're Training Your Triceps All Wrong!
Recommendation · BodyChoose exercises that put the shoulder in flexion and the elbow up overhead to stretch the long head of the triceps.
ATHLEAN-X™ · STOP, You're Training Your Triceps All Wrong!
Creator perspective · MoneyKatie describes days 31-60 as taking all the information gathered in the first 30 days and asking: what is actually going on here, what needs to happen to move the needle, and how am I going to move the needle. You better have a hypothesis and a plan built on how to execute it. By 60 days, the plan better be in place, cranking, and going immediately — or else you won't have a job.
30 Minutes to President's Club | No-Nonsense Sales · #622: Sales Enablement Playbook: How To Win Your First 90 Days
Creator perspective · MoneyThe other half of the first 30 days is the revenue velocity model, where you pull metrics to help inform what actually needs your time versus what might be moderately painful for people but isn't actually going to move the needle.
hedged
30 Minutes to President's Club | No-Nonsense Sales · #622: Sales Enablement Playbook: How To Win Your First 90 Days
Creator perspectiveWhen a customer reports a critical failure you can't reproduce (because you can't see their real data), ask them to anonymize the documents and problem so you can recreate and fix it.
Latent Space · Building Legal Agents When You Can't See the Data - Jacob Lauritzen, CTO of Legora
ReportLegora fine-tunes models only for specific bounded tasks where a post-trained model offers a large delta over frontier models in cost and latency, and does not attempt to train its own main agent model.
Latent Space · Building Legal Agents When You Can't See the Data - Jacob Lauritzen, CTO of Legora
Creator perspectiveDon't commit to a public API surface while the product moves fast and the team is small — maintaining old API versions for a changed data model slows the team down too much. Delay offering a full API surface until the company is big enough to absorb the maintenance cost; at 20 engineers it's not viable, but at larger scale it becomes a cost you can take on. Legora started offering APIs only about six months ago, after a long period of not having them.
Latent Space · Building Legal Agents When You Can't See the Data - Jacob Lauritzen, CTO of Legora
Creator perspectiveJacob Lauritzen explains Legora cannot look at any customer traces or data because all legal data is super confidential — much of it belongs to the law firm's clients, not the firm — making it extremely frustrating to understand what users actually do. A dedicated evals team approximates what customers are doing and builds those approximations into eval sets.
Latent Space · Building Legal Agents When You Can't See the Data - Jacob Lauritzen, CTO of Legora
Creator perspectiveFine-tune models only for narrow, bounded tasks where there is a large delta between frontier model cost/latency and what a post-trained model can achieve. Use post-trained small models for legal research retrieval — fan-out retrieval queries can be executed super quickly, cheaply, and at low latency. Don't try to train your own general agent to compete with frontier labs — you won't win against Anthropic.
Latent Space · Building Legal Agents When You Can't See the Data - Jacob Lauritzen, CTO of Legora
Creator perspectiveBuild a multi-layer citation system where every agent output links to a source — clicking highlights the exact location (bounding box in a PDF, cell/row in a table) so users can validate claims multiple layers down.
Latent Space · Building Legal Agents When You Can't See the Data - Jacob Lauritzen, CTO of Legora
ReportManish demonstrated that the code layer running on the Figma canvas can be mirrored to a phone, where it runs responsively — enabling on-device testing of code prototypes.
Figma · Config India 2026: Figma deep dive: Design together with code on the canvas
Report · MoneyContinuum can be connected to a coding agent so the agent can use it to visualize the user's problem — an underexplored capability called 'visual thinking'. Instead of a coding model thinking through a problem only via code, the model can spin up a version of Continuum to think through the problem visually (e.g., creating a 3D model of a bike). Agents have the ability to do visual thinking — spinning up 'imagination' that users can tap into and interact with. Users can watch the agent's visual thinking themselves and guide the agent in better ways than guiding it via text alone. Visual thinking is described as one of the most underexplored capabilities demonstrated, and guiding an agent through shared visual thinking is framed as a completely new paradigm.
hedged
AI Engineer · You Can't Tell What's Real Anymore. Here's What Comes Next - Cristóbal Valenzuela, Runway
Principle · MoneyLenny describes a framework he loves: the human comes up with the idea and direction at the top, AI does a bunch of the work in the middle, and the human reviews and iterates until it's done — the AI sandwich or human sandwich.
Lenny's Podcast · Molly Graham: The grief, burnout, and opportunity hiding inside the AI transition