brian

My ldl-c went from 120-130mg/dl to 70-80 mg/dl with diet alone in 2015. Graph below suggests this puts me into the range where you start to see a small amount of regression or at least slowing/stabilisation of plaque. I’ve had a repeat CT angiogram and am having HeartFlow analyse my baseline and follow up scan. And I’ll have an episode out soon discussing the results with 1-2 cardiology/lipid/cardiac imaging trained doctors.

Sleeping together is Not love. Talking 24/7 is NOT love. Laying up is NOT love. Going on dates is NOT love. Love is someone seeing the absolute worst side of you & still loving you the same. It's someone making stuff happen for you when you can't. It's someone holding you & calming you down while you're pouring your eyes out. It's someone voicing all the good things about you when all you see is bad. It's someone making sure you're okay, someone praying for you every day & night. It's someone that makes you laugh/ smile when you feel down.

One of the very best PhD students I worked with (very successful prof now) worked 7am-3pm, didn't chit-chat, & did stunning work: much like this post. I was the polar opposite. I worked out okay too, and benefited from the lollygagging. Different people, different models, no ∀.

@theproof function health?

Sperm check ✔️ Video to follow @theproofwithsimonhill?si=r5NYMlXOvPuxocXS" target="_blank" rel="nofollow noopener">youtube.com/@theproofwiths…

#setthecurve they hate me

It is an absolute amazing environment if you are a derivatives TRADER. Bubble, no bubble, macro, whatever narratives, idk and idc. The market is handing out opportunities all day long so forget everything else and make hay while the sun shines 🙌🏾.

The AHA just published a peer-reviewed review asking whether lean keto dieters with LDL-C spikes should be worried. Spoiler alert: The answer in the paper: yes. DiMattia and Petersen in JAHA today: normal-BMI adults on ketogenic diets show mean LDL-C increases of 18 to 70 mg/dL, with ApoB rising 35 mg/dL alongside. The effect is substantially larger in lean individuals than in those with overweight or obesity, where LDL-C either does not change or falls. The mechanism: reduced hepatic cholesterol synthesis in lean individuals increases sensitivity to dietary cholesterol and suppresses LDL receptor clearance. The lean mass hyper-responder phenotype gets a full clinical discussion. The authors state directly that the claim of cardiovascular benignity for this lipid triad has no rigorous supporting data. The KETO-CTA preprint data cited: 85% of participants showed plaque progression, 37% relative increase in noncalcified plaque volume over one year. The ApoB null finding in that dataset is explained: the study was not designed or powered to evaluate that relationship. No long-term cardiovascular outcome trials exist for ketogenic diets in lean individuals. The paper states this gap explicitly: absence of outcome data is not evidence of cardiovascular safety. Plaque is a surrogate. MACE is the question. That question has not been answered. Carbohydrate reintroduction reverses the LDL-C elevation. Pharmacologic therapy adds further reduction. The clinical path forward exists. The evidence of safety does not. Isn't it ironic. A term coined by a citizen scientist to describe a phenotype proposed to be benign now anchors the clinical challenge section of a peer-reviewed AHA journal. The evidence that phenotype is safe is precisely what the paper says is missing. doi.org/10.1161/JAHA.1…

det( λ Ι - A )

as my final act of failure i'll do masters

One theorem every ML engineer should know: The Johnson–Lindenstrauss Lemma. It states that high-dimensional data can be projected into a much lower-dimensional space while approximately preserving pairwise distances. Why it matters: • Explains why random projections work • Enables scalable learning in high dimensions • Used in embeddings, compressed learning, and ANN search • Helps fight the curse of dimensionality The surprising part: You can reduce dimensions dramatically without destroying the geometry of the data. That’s why many ML systems can operate efficiently even with massive feature spaces. Modern representation learning is deeply connected to this idea: Good embeddings preserve structure while compressing information. In ML, compression is often not loss of intelligence — it’s removal of redundancy.

A single 25 mg dose of psilocybin leads to brain structural changes that were seen at 1 month. From a cross-over study of 28 healthy volunteers, no prior psychedelic, who also were also assessed after 1 mg. Behavioral results in Figure nature.com/articles/s4146…

In this episode, I sit down with Dr Andrea Glenn, an NYU professor and Harvard visiting scientist, to look carefully at what the evidence actually says about plant-forward eating, the Portfolio Diet, butter versus plant oils, and the protein conversation almost everyone is missing the point on. For the full show notes head to: theproof.com/%e2%81%a0the-p…

The honor system is over. Apple now wants to see some ID before handing out the student discount.

“the hantavirus kills you too effectively for it to become a full blown pandemic” is the kind of jaded analysis I look for from a virologist

New Anthropic research: Natural Language Autoencoders. Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read. Here, we train Claude to translate its activations into human-readable text.

Terence Tao is answering a fundamental question regarding the safety and reliability of modern AI: "How can we use a tool that is powerful, but unreliable?" W = ∑(wᵢ ⋅ xᵢ) + b AI isn’t just about “smart”; it’s about the probability of *looking* right. We’ve built systems where the weights (wᵢ) are optimized for plausibility, not veracity. This creates a “convincing mirror” that confidently serves dangerous advice in medicine or finance. The gap between “convincing” and “correct” is the most critical variable we need to solve for.

@adnauseams babe wake up, new food daily xime post

grass-fed cheddar, farm eggs, ancestral blend ground beef, local sourdough with avocado, microgreens and pumpkin seeds, and organic morning blend juice with raw honeycomb

@adnauseams i fw optics

physics lecture final :p

How can you be bored when you’ve got a career to build,a body to perfect,a mind to make smarter and success to chase.

@certifiedsauce1 2mg

How much GHK-cu peptide are yall taking subq?