Principle · BodySome athletes have tons of strength yet seem to collapse at high speed because they only have strength when they have time to express it. When contact time is constrained, maximal strength gives you nothing, and vice versa. Usually, when you have a short contact time for a high flight time, you tend to be more stiff, and a huge collapse or trunk movement at initial contact indicates an issue with reaction and very fast actions.
Garage Strength · Why Most Athletes Train Speed Wrong w/ Sprint Scientist Dr. Jean-Benoit Morin
Creator perspective · BodyDr. Morin explains the paradox of athletes with high maximum strength (e.g., on plantar flexion iso tests) who still collapse at high speed: they only have strength when they have time to express it, and when time is constrained they have nothing.
Garage Strength · Why Most Athletes Train Speed Wrong w/ Sprint Scientist Dr. Jean-Benoit Morin
Principle · BodyThe goal of sprint exposure programming is to avoid both underexposing and overexposing athletes, since these are two almost equal types of risk. He mostly sees players who are underloaded rather than overloaded.
Garage Strength · Why Most Athletes Train Speed Wrong w/ Sprint Scientist Dr. Jean-Benoit Morin
Principle · BodyThe goal with max velocity sprinting exposure is to avoid both underexposing and overexposing athletes: underexposure and overexposure to high-speed running are two almost equal types of injury risk. Players who get less than the recommended high-speed running volume seem to have problems more often, and players who get more must be really adapted or it becomes too much. Doing too much top speed sprinting exposes players to overexposure and possible injury risk.
Consequential creator advice requires independent verification before Today can promote it.
Garage Strength · Why Most Athletes Train Speed Wrong w/ Sprint Scientist Dr. Jean-Benoit Morin
Creator perspectiveDepending on how big the workloads are, you're going to have different bottlenecks and challenges with distributed sandbox setups. When running only a couple of sandboxes, the problems are individual sandbox failures, like why one or two sandboxes are breaking. Once you're dealing with many sandboxes at scale, the problems shift to managing logging overhead, tracking sandboxes, and keeping track of the network on the machines they run on. Each machine can only talk to so many sandboxes at a time, which is a scaling constraint.
AI Engineer · How I Learned to Stop Worrying and Love the Sandbox — Matt Brockman, E2B
Creator perspectiveRunning AI coding agents locally on your own computer risks the agent using up all your CPU, spawning processes, deleting files, and accessing environmental variables. Running them on a shared server exposes your secrets and environment to the agent. AI-written code can spawn processes that consume excessive CPU. Sandboxes isolate these risks.
AI Engineer · How I Learned to Stop Worrying and Love the Sandbox — Matt Brockman, E2B
Principle · MoneyIf an agent achieves its goal using a skill, the skill works and becomes more trustworthy; if it fails, the skill becomes more robust because the gap is now known and can be updated.
AI Engineer · From 36% to 100%: How Self-Improving Agents Write Their Own Skills — Rafal Wilinski, Runlayer
Creator perspective · MoneyDeploying agents without company-specific context and playbooks leads to costly failures, because agents try to rediscover embedded knowledge from scratch and often fail.
AI Engineer · From 36% to 100%: How Self-Improving Agents Write Their Own Skills — Rafal Wilinski, Runlayer
Principle · MoneyA skill is a playbook an agent can read when it judges the knowledge relevant. Technically this is progressive disclosure: the harness first sees only a short summary, then reads the full skill and injects it into the main context when useful. Injecting skill knowledge can change the agent's trajectory, and skills are a way to retain procedures, improve agents, and guide models.
AI Engineer · From 36% to 100%: How Self-Improving Agents Write Their Own Skills — Rafal Wilinski, Runlayer
Creator perspective · MoneyLetting a long-horizon agent reason without an initial trajectory guide causes it to dig deep rabbit holes, spending hundreds of tool calls defending a wrong thesis and wasting time, tokens, and money.
AI Engineer · From 36% to 100%: How Self-Improving Agents Write Their Own Skills — Rafal Wilinski, Runlayer
Principle · MoneyModels are becoming a commodity everyone has access to, and the cost of software is going to zero, so anyone can rebuild your product at least on the surface. Your moat can be building the self-improving flywheel and distilling knowledge through it, because it captures your own hard-won lessons — governed, compounding, raising the floor for everyone across the company. Building the knowledge flywheel is also a way to be resistant against model intelligence deprecation, and distilling skills is a great way to lower your costs without changing the model. One thing a competitor can't copy is your internal culture and your identity.
Commercial context: The creator describes this as their own product.
AI Engineer · From 36% to 100%: How Self-Improving Agents Write Their Own Skills — Rafal Wilinski, Runlayer
Principle · MoneyYou should be distilling skills because frontier intelligence is rented or borrowed and can disappear — it can deprecate or get dumber, and people complain daily on Twitter and Slack that Opus is really dumb today. While you have access to the best version of frontier intelligence, use it to discover and persist useful procedures. Build a catalog of how the work gets done inside your company as a collective subconscious.
AI Engineer · From 36% to 100%: How Self-Improving Agents Write Their Own Skills — Rafal Wilinski, Runlayer
Principle · MoneyThe core idea to carry forward: exploration should be turned into reusable procedural knowledge. When an agent cracks a genuinely hard problem, write down exactly how it was solved and hand the playbook to the team — like the best engineers, salespeople, and lawyers who document how they cracked hard problems so others don't have to figure it out from scratch. Otherwise the knowledge, debugging, and reasoning behind a cracked problem is just gone when you move on to the next task or close the session.
AI Engineer · From 36% to 100%: How Self-Improving Agents Write Their Own Skills — Rafal Wilinski, Runlayer
Creator perspectiveDuncan Trussell argues that fixating your attention on the state or political news is a form of idolatry: you begin worshiping the state and the story in the news. Whether you hate or love the state, the Dems, or the Republicans, all of your attention is fixated on that thing — and in bhakti yoga, where attention goes, energy flows. Via fixation, the rest of reality dims until you're just saying a mantra against or for the state, and you become a cultist engrossed in a new religion of the state via hate or adoration — it's the same thing. Don't trick yourself into thinking that because you hate Donald Trump you aren't bowing down at that altar.
spiritual/philosophical interpretation offered by the guest
PowerfulJRE · Joe Rogan Experience #2391 - Duncan Trussell
Creator perspectiveJoe Rogan argues people are waiting for a top-down solution instead of acting themselves, and that anyone could drive a truck to a dumpster and clean up trash rather than waiting for the federal government.
PowerfulJRE · Joe Rogan Experience #2391 - Duncan Trussell
Principle · MindSociety overemphasizes optionality, which is really a form of delaying gratification; options only have value when exercised, and dying with unexercised options is like an ant dying next to a pile of crumbs it never ate.
The ant analogy is explicitly hypothetical.
Alex Hormozi · It took me 37 years to realize what I'll tell you in 18 minutes.
Principle · MindDivide your day assuming roughly 12 usable focused hours after ~4 hours for eating, showering, hygiene, and living, then allocate across your ranked priorities. To improve a neglected category, make explicit trades — decrease investment in other categories to make room rather than trying to add without subtracting.
Alex Hormozi · It took me 37 years to realize what I'll tell you in 18 minutes.
Principle · MindThe goal of youth is optionality and the goal of adulthood is commitment, and the two are at odds; at some point you must reconcile which you are, because being an adult means eliminating alternatives and walking through one-way doors you cannot walk back through.
Stated as the speaker's view.
Alex Hormozi · It took me 37 years to realize what I'll tell you in 18 minutes.
Principle · MindThe root of dissatisfaction is wanting it all: we want the benefits of trades we did not make plus the benefits of the ones we did. At the end of life you may wish you had done the other things, but you say that from the position of having already earned the benefit of your existing trade. The speaker's takeaway when reasoning to its natural end: 'I am the problem.'
The 'I am the problem' line is the speaker's personal reflection.
Alex Hormozi · It took me 37 years to realize what I'll tell you in 18 minutes.
Principle · MindIf you like the priorities that you have, you have already won. People get in trouble when they allow life to make their priorities for them instead of making the choices themselves, and sitting passively by while options expire is a failure mode.
Stated as the speaker's emphatic view.
Alex Hormozi · It took me 37 years to realize what I'll tell you in 18 minutes.