36,915 results · Most relevant
Prompt the agent to apply your library to a screen and observe whether it reaches for your components, uses your tokens, and follows your conventions without them being included in the prompt.
Eugene Teo advises the host's team to cut posting frequency, arguing that a small number of posts vastly outperform the rest, so the time spent on low performers should go into making more of the exceptional ones.
advice attributed to Eugene Teo; based on the team's own performance data
Eugene Teo recommends hiring a videographer to follow you while training athletes and clipping the best raw moments for social media, posting raw unstructured fly-on-the-wall training footage rather than super hype Mr Beast-style edits, and publishing one or two training sessions with an athlete each week as long-form content.
Eugene Teo advises the host to experiment with increasing high-TAM content volume in the short term, since the host hasn't done much of it yet.
Eugene Teo recommends inviting social media creators with their own brands and audiences to your facility to train and film the session, filming a live assessment session putting a guest through your periodic testing battery (e.g., the 13 tests done every 13 weeks) and showing the results, and being prepared to invest financially in creator collaborations — flying them out and taking good care of them so they have an awesome experience.
Huberman recommends deliberately choosing your eating condition for each focus bout — e.g., fasted for morning work, fed for afternoon — and scheduling your hardest focus work early in the day. He notes that after eating he personally experiences diminished focus capacity, so hard work should come before meals, and if you wake up very hungry, eat earlier rather than forcing a fasted session.
Hormozi advises that if you enjoy competing and like having the target, control the controllables: look the way you want to look, work your ass off on your posing, prep the way that you know that you should prep, then go in and do your best.
As a developer, secure your sandboxes to prevent AI agents from accessing external APIs or infrastructure to acquire resources (e.g., task budget, compute) for goal completion.
Contested by Dru Knox on AI Engineer, Matt Brockman on AI Engineer
Implement multi-layered mitigations including fallbacks, classifiers, and constitutional guidelines, as single points of failure are insufficient against sophisticated agent exploitation; Anthropic has many mitigations at many points.
Going a little wider with the grip on a low bar squat allows more elbow drift upward, moving out of some external rotation into a bit more internal rotation to alleviate discomfort. If that doesn't bring relief, the creator advises switching to a high bar squat with the bar on the upper shelf of the traps, which is more forgiving from a flexibility standpoint. A front squat is also an option because it has none of the external rotation demands of the low bar position.
If discomfort persists after making bench press tweaks, the creator advises skipping the barbell bench press and shifting to dumbbells. When lowering the dumbbells, go slow to make sure nothing is being aggravated, because going fast removes control. Regardless of how heavy the dumbbells are, going slow with elbows tucked and a good wrist-to-elbow relationship allows continuing dumbbell bench presses without stopping altogether.
The creator advises stopping preparation an hour before a high-pressure event and instead priming yourself for comfort, doing something beforehand that gets you into the state you want to embody in that situation.
If your business is at a place where you can do it, hire team members with different motivation archetypes than your own — if you're a craftsperson, you need to hire a people person; if you're a people person, maybe hire a list maker. Bringing in team members with different motivational profiles offsets what the company needs so you don't have to do all the things yourself.
Anish Acharya, speaking from his VC perspective, says investors would happily bet on a product without a defined moat or durability story if it has a lot of momentum, a lot of craft, and growing engagement. He recommends founders focus on these rather than a moat narrative.
Anish Acharya recommends paying for the highest intelligence — expensive but performant frontier tokens — for job functions with unbounded upside like drug discovery, sales, support, research, and engineering. For functions with bounded upside like legal and finance, he recommends Pareto-efficient, mid-tier models that deliver good performance at a good price.
Consequential creator advice requires independent verification before Today can promote it.
Anish Acharya recommends building the mental habit of asking 'How do I have Claude do this thing for me?' as a way to offload tasks to AI. He presents it as a script everyone should internalize to reframe everyday work through an AI-assistance lens.
Anish Acharya describes distilling a go-to-market team's job down to the core human skills they're best at — like relationship-building and closing — while AI handles all the administration around that core work. Not doing this leads to wasted time on non-core work.
Anish Acharya argues that company building is increasingly a matter of creating loops — sets of agents doing tasks — that can run autonomously, from loops per person up to loops that run large parts of the company.
Coach Kyle advises reopening by commenting on something in the environment (e.g., 'the sandwiches look so good') and acting as if the conversation is still going so it doesn't get weird, rather than asking 'is it cool if I talk to you for a second.'
Coach Kyle advises that when asked a question like 'what are you,' you don't always need to just answer in a boring way — instead, put it back on her, have a little fun with it, and make her guess by making an assumption about what ethnicity you think she is.
Choose sources → to get a library of your own and a daily brief.