35,800 results · Most relevant
Dr. Scott Forbes explains that with creatine, you may be able to do one or two extra reps per set in the gym. Training a little harder with one or two more reps per set over time makes you bigger and stronger — this is the main theory of how creatine works. Retaining muscle as you age helps maintain muscular strength.
Dr. Forbes explains there was no creatine hair loss study until last year, when he, Dr. Jose Antonio, and Grant Tinsley ran one; the 12-week study with a hair specialist found no detrimental effect on hair health or quality, though genetic factors and predisposed individuals were not covered.
timeline flagged as possible numeric error; limitations on genetics noted
Dr. Forbes describes roughly four or five studies showing creatine improves brain performance when the brain is stressed, such as under sleep deprivation, and recommends it for cognition after bad sleep, for a surgeon up all night, and for time-zone travel.
study count is approximate ('maybe four or five'); studies used severe sleep deprivation
Coach Kyle argues that winning over the friends or family of a woman you approach is really important in any environment — nighttime, social circle, house party, or day-to-day — because when you leave, they all start talking about you, and that talk influences her attraction toward you.
Bridgewater built The Vault, a data platform that turns unstructured data into a first-class asset in their investment process.
The five dimensions that matter when building for humans and agents are recall, precision, filtering, entitlements, and scale.
A technology doesn't need to be better than the last thing at everything — it just needs one feature that makes a breakthrough. The smartphone was panned as a bad phone and a worse computer, but its camera and GPS in your pocket enabled entirely new applications like Uber that a PC could never support.
Ben Horowitz says the doubts are going to be there — ignore the doubts. Technology can't automate maintaining confidence through uncertainty and continuing to bet on yourself when the doubts inevitably show up.
Ben Horowitz says you don't take out a competitor with an equally good or slightly better product — you need something like a 10 times better product; if you take the market, that's probably the most obvious, most significant thing.
hedged
Ben Horowitz argues that proprietary data generated from serving high-end users — e.g., talking to super high-end programmers all day — becomes a key differentiator for AI applications, and that winning the market compounds defensibility because you get more data about your particular kind of user.
Andrey Rybka argues that thinking 'we just need a better sandbox' is not a sufficient security strategy for AI agents, since agents are only useful if they can reach many tools, services, and data — and if given that reach with no hard limits, the damage could be unlimited.
When an agent is given a goal and broad access to tools, services, and data with no hard limits, a hijacked goal can cause unlimited damage. The speaker argues that firewalls and static role-based controls cannot detect hijacked goals, and that deterministic controls outside the model checking every call are needed instead.
The speaker argues that an agent's acknowledgment of an instruction (e.g., 'understood, thumbs up') is not trustworthy evidence of compliance and must not be relied on as a control; deterministic controls, circuit breakers, human approval, scoped credentials, and tenant isolation are what actually prevent damage.
Andrey Rybka states that input outside of context cannot give commands — data must be separated from instructions.
There is no inherently good or bad association — it's only good or bad relative to your ideal customer (ICP/avatar). If your ideal customer likes an association, some of those cool points rub off on you. After a hundred cool associations, you can appear without the association and some of the coolness still rubs off — until eventually you become a brand.
The third and final step of building a brand is 'defending the belt.' Once you get to the top, you have to stay relevant because that's what degrades large brand empires. Big brands get nibbled away into irrelevance as smaller players do things that are more relevant for smaller audiences and start gaining market share. A big brand has to fight newer, hungrier challengers to defend the belt. Somebody can't just win the UFC and then say 'I've won forever' — they've got to stay in the title spot and let people take shots at them.
Defending the belt means doing an aspirational thing that your audience finds cooler. Big brands have to show people what's up — remind them who's boss and who can still show up. Can John Jones still show up and still put a hurting on someone — the question a champion brand must keep answering. You have to have these belt moments.
Once you have the credibility, associate it with yourself. In practice, this means associating yourself with positive things over and over again. 'Positive' is defined by what your audience or ideal customer finds positive — which means you have to know who you're going after.
The first stage of building a personal brand is to accomplish real things first — 'do epic shit.' Everyone skips this step because it is the hardest, but an epic personal brand requires doing epic things first; the proof is in the pudding.
New compute market structures are needed mainly to solve mismatches: bridging capacity, more flexible commitments, resale rights, and protection against price moves. From an inference cloud point of view, selling a fixed-price service while the GPU bill floats creates price exposure. Locking capacity for years can solve part of the price exposure problem but creates idle-hour exposure elsewhere unless you're always using the capacity. Reselling spare capacity only helps if someone can actually buy the hours. A financial hedge can address a reference price but cannot deliver the cluster right now in terms of cash settlement or guarantee utilization. The useful structure has to start with the particular risk that the business has and needs to manage.
Choose sources → to get a library of your own and a daily brief.