Patrick Samy

In less than 5 years, the market cap of @ouraring, @whoop, @levels, @eightsleep, @span_health will be 10x @fitbit today.

Dad Mode shares a lot of similar skills with Founder Mode.

@tjparker 💚

I think my many-thousand-year-old tree, which was transplanted into the courtyard at my house 18 months ago, is finally starting to push out new growth. And boy, does that make me happy.

Couldn't agree more: this will be the most impactful impact of AI.

There's only a handful of products you can design that will change millions of lives for the better. This is one of them. Whether you work on a consumer product in big tech or startups, if you have exceptional attention to detail and live to design unique experiences and category-defining products, we want to have a chat! Location: Europe or US East Coast

@HalePrevention I need an appointment this week. Is that possible?

Get your Calcium Score today. Best predictor of heart disease. - $249. No referral needed. - 10 minutes per scan. - Same day appointments. - Live across all 50 states.

@sama I do that a lot. Also to chat with past context and remember very specific details and decisions.

people are really starting to use voice to interact with AI, especially when they have a lot of context to dump. GPT-Realtime-2 comes to the API today; it is a pretty big step forward. (we are working on improvements to voice in chat.)

@moguzbulbul Followed you back! Or email me at patrick@joinhale.com

@patricksamy Would love to talk interfacer.co

I’m hiring a Founding Designer to build the future of preventive health at Hale. We're a small, low ego, high ambition team, building a company where smart generalists do the most important work of their careers and have fun doing it.

@robably__ Awesome! Can you email me at patrick@joinhale.com?

@patricksamy I'm down to chat robably.net

@ammaddyaseen Followed you back! Email me at patrick@joinhale.com

@patricksamy Hey! I’m a designer and really like what you’re building around preventive health at Hale. I’d be interested to learn more about the Founding Designer role. Is the role remote onsite or hybrid? Your DMs seem closed on my side can you open them?

If this sounds like you, DM me or email patrick@joinhale.com

🤖 You love contributing code directly with AI when needed 😍 You pay attention to details and care about delivering world-class UX 🦄 Your past teammates and managers say great things about you 🚀 Location: ideally UK, open to EU and US East Coast.

Vibeware is the new software.

@HalePrevention Do you do same week appointments in CA?

Get your Calcium Score today. Best predictor of heart disease. - $249. No referral needed. - 15 minutes per scan. - Same day appointments. - Live across all 50 states.

@zehranaqvi Legendary launch video 🔥

The internet curated for the obsessed. Welcome to Lore.

Launch it now 🚀 @vonderleyen

A subscription product with no tangible recurring value is destined to fail. Consumers are also pretty smart in their perception of tangibility.

Ilya said the quiet part out loud on Dwarkesh's pod, but most people still aren't processing what it means. Here's what's actually happening inside AI labs. Research teams have entire divisions that do nothing but create new RL training environments specifically designed to boost benchmark scores. They treat AIME, SWE-bench, and MMLU like standardized tests. The model practices 10,000 hours on competitive programming problems until every proof technique is at its fingertips. Then it fails to fix a simple bug in production without introducing two new ones. Sutskever used the perfect analogy. Student A grinds 10,000 hours of competitive programming. Memorizes every algorithm, every edge case, every proof technique. Becomes the #1 ranked competitive coder in the world. Student B practices 100 hours but has "it." Intuition. Taste. The ability to learn new things quickly. Who has the better career? Student B. Current AI models are all Student A. The benchmark gaming runs deeper than most realize. Studies have shown data contamination inflates model scores by 20-80% on popular benchmarks. The training-test boundary is porous. Models memorize answers rather than learn concepts. And when you control for contamination, much of what looks like intelligence is pattern-matching on seen data. This explains the economic puzzle Ilya pointed to. Models score 100% on AIME 2025. They hit 70%+ on GDPval beating human professionals. Yet businesses still struggle to extract value. The benchmark performance says genius. The P&L says otherwise. The sample efficiency gap tells you everything. A human teenager learns to drive any car after 10 hours. An AI model might need millions of examples and still fail on slight variations. A human learns a concept once and applies it everywhere. Models need to see the exact pattern thousands of times and still choke when the formatting changes slightly. Sutskever's diagnosis: we're moving from the "age of scaling" (2020-2025) back to the "age of research." The belief that 100x more compute would transform everything is dying. His $3B company SSI is betting that the next breakthrough comes from solving generalization, not stacking more GPUs. The labs know this. That's why the benchmark arms race is accelerating. It's easier to show impressive numbers than admit the fundamental approach might be plateauing.