Jolly

@kavi_mathur @cremieuxrecueil Believable ;)

@cremieuxrecueil How does 8sleep compare?

Compared to actual polysomnography, it seems even the best sleep trackers aren't that accurate when it comes to measuring sleep stages.

@reggylong lets talk

Do you want to: 1. Go deep on sleep and read tons of research papers. 2. Meet the worlds leading sleep experts and grok the ideas better than anyone else. 3. Do nothing but read and write all day long. Reach out. We have a dream-job.

@TylerAlterman @reggylong oh and you might like my diy navage cpap mask

@TylerAlterman @reggylong i've been geeking out on sleep so much lately...

Proof that you can just do things - I am working on a cure to my own sleep apnea. After 11 different treatments (total of ~$100K), doctors told me I'd exhausted all options. I knew there was more to be done. If you have brain fog, feel tired all day, DM me - I can help.

@bitcoindata21 A lot of your newer posts just read like AI nonsense optimized for views/shares, and not the sort of detailed analysis you used to post.

Apologies to everyone about what I've been showing over the last few months. It hasn't been good enough. Hindsight is annoying. Even from what I have gotten on top of now, if I had got things right I would have sold between 110-120k and bought 80-90k. That said, I still believe this is an intracycle reset like march 2020 or April 2025 (from a business cycle perspective). This is why I still don't align with the end of year bottom and/or target of 40-50k. If I did believe it was end of business cycle, I'd be looking for that area. I will be aiming to consume/post 50-75% less on social media, so I can really focus properly to staying on top of things to a better level.

@NukitToBeSure Does the noise change if you size up to use 3-5" thick filters instead since pressure drop improves? Or is that improvement just seen in air flow?

Spoiler- the included custom fans underperform Arctics, but they are quiet and very, very inexpensive, so they cost almost nothing to include. Most of the Tempest Pro's cost is shipping, to be honest. The steel is very heavy, but it should last virtually forever, and I don't see HVAC filters or PC fans becoming obsolete in the next few decades (we still use E26-threaded light bulbs, and those are 140 years old). A lot of buyers wanted an "all-in-one" kit, but including Arctic P140 PWM would have meant a lot of wasted money for the 20-30% of buyers who are hardcore customizers that wanted Arctic P140 Pros or RGB fans. This was the solution- fans included, so you can get up and running, but low-cost enough that if you want to customize, it's not a lot of wasted money. A lot of PC filter-box designs are very North America-centric, when HVAC filters are hard to obtain in other countries. We went with 1x20x25 because that is the size that is easiest to obtain internationally- Digikey has it. The main complaint about the more efficient 2x20x20 the Original Tempest uses is that many people couldn't get it at their local hardware store, even in the US. This led to problems as buyers ignored repeated warnings and instructions, tried to stuff 1x20x20 filters in (which are a larger nominal size), and then wanted a refund after that did not work. There are lots of great options out there; the Tempest Pro is just one for people who want something wall- or ceiling-mountable and customizable that lasts indefinitely.

@open_erv You could be in Puerto Rico, my indoor is 65% after AC's and a dehumidifier

The relative humidity in my home right now is 19%. A little higher than it was during the day in the Sahara yesterday cliffmass.blogspot.com/2023/10/the-hu… wunderground.com/hourly/cf/saha… The optimal is 30-50%. My humidifier only increases this a few percent, and uses 4 liters of water a day. It takes a lot of energy to humidify all that air optimally. This is why I need to get some silica gel on the heat exchanger in my energy recovery ventilator. Last year I tried this, and it helped much more than the humidifier, but the adhesive began to smell significantly bad. I have found a new non-stinky and otherwise suitable adhesive but have yet to fabricate some heat exchangers with it.

@p_ezravasquez @SwipeWright Qualitative != Quantitative. What's your point?

I'm in an education PhD program right now and this is from the textbook on qualitative research: Source: Ravitch & Mittenfelner (2021). Qualitative Research, 2nd Ed. Sage.

A new paper in the journal Sociology laments the difficulty in teaching quantitative research methods to sociology students because many have a negative "attitude toward numbers" and even find numbers "triggering." Perhaps one way sociology might attract more intelligent students could be to stop using the word "fuck" in its paper titles.

@KateD_AI i hope you are doing ok and the surgery went well!

And the chance of failure during surgery is like not 0 so I’m just trying to hype myself up as much as possible. Thanks for listening and all your well wishes.

I’m 4 days out from my surgery now, wanted to share a little bit more (as much as I can without violating my NDA) Basically my whole life I’ve had schizophrenia that caused me to suffer nervous breakdowns due to paranoid delusions and visual/auditory hallucinations. (THREAD)

@dguido Did you give it direct access to raw DNA data or only synthesized reports? Any plans to upload your rules/prompt?

Here are some examples of my CLAUDE rules. I used the Scientific Skills from K Dense AI (github.com/K-Dense-AI/cla…). I also wrote my own lab-report-converter skill using Mistral OCR.

I used Claude Code to analyze my health records. 2 days of work resulted in 16 changes to how I'm taking care of myself. It feels like cheating. This is the most in-depth and precise analysis of my health I've ever done. I built a personal EHR: I archived 6 years of lab results, visit summaries, and genetic data, converted it all to markdown, then added my current supplements and lifestyle info. I'd been monitoring my labs for years but mostly ignored my genetic results (Sequencing, StrateGene, RootsGA, Nebula). With Claude, it finally felt achievable to dig in and see if the genetics explained my treatment outcomes and bloodwork patterns. I also took a renewed look at all my existing treatment plans and found a bunch of potential improvements. I'm fortunate to work with excellent doctors (The Lanby, Central Park Endocrinology, Mt Sinai, Ember Health). Claude didn't replace them. It helped me interpret what they told me, expand on their recommendations, and understand what else I could do. My doctor's visit summaries were key inputs and I'm reviewing all these changes with them. I gave Claude interpretation rules: genetics are risk modifiers, not destiny; prioritize silent killers like cardiovascular and cognitive decline; verify claims against real databases. I wired in ClinVar for variant pathogenicity, DrugBank for supplement interactions, and PubMed for evidence quality using skills from K Dense AI. A note on percentiles: polygenic scores show where I fall in population risk distributions, not certainties. 98th percentile means higher risk than 98% of people tested, not a 98% chance of having the condition. Here's everything I was able to discover by assembling the right context, providing the right rules, and prompting Claude Code over about 2 days: APOE ε4: Brain & Heart I carry APOE ε4 (rs429358 T/C), roughly 3x risk for late-onset Alzheimer's. I also have 100th percentile polygenic burden for ApoB and LDL (Richardson 2020, PLoS Medicine, 440K individuals, 255 variants). Standard lipid panels underestimate my cardiovascular risk. I ordered ApoB testing for January. CDP Choline: My body makes 30% less of a compound called phosphatidylcholine (PEMT rs7946 TT). Phosphatidylcholine is a building block for brain cell membranes, and it's also what your body uses to make acetylcholine, the neurotransmitter involved in memory. With APOE ε4 already putting my brain health at risk, I didn't want a second strike against it. CDP-choline gives my body the raw material to make phosphatidylcholine through an alternative pathway. I added 250mg daily. Vitamin D + K: I learned that vitamin K2 directs calcium to bones and away from arteries (a mechanism that's promising but not yet proven in trials). With my APOE ε4 cardiovascular risk, adding K2 made sense. I swapped my weekly calcifediol for D3 + K2 3x/week. Glutathione: I have two genetic variants that affect how my body handles toxins and oxidative stress. The first (GSTP1 rs1695) slows down a detox enzyme that needs glutathione to do its job. Supplementing glutathione gives that enzyme more of what it needs to work. The second (PON1 rs662 TT) reduces my protection against oxidized LDL, which is the form of cholesterol that actually drives plaque buildup in arteries. Glutathione is a general antioxidant that may help reduce oxidative damage. I added 400mg daily to address both. Melatonin-SR: I have 92nd percentile genetic risk for insomnia (Lane 2019, Nature Genetics, 57 variants including MEIS1) and a lifetime of sleep difficulty. With APOE ε4, sleep quality is neuroprotective. During deep sleep, the brain's glymphatic system clears amyloid, the protein that builds up in Alzheimer's. I added 3mg sustained-release melatonin for 6-hour coverage. Septoplasty Upgraded: I've always had chronic congestion and trouble breathing at night. Genetics revealed why it's severe: when I have an allergic reaction, it hits hard and lingers longer than normal (HRH1 fast histamine response, MAOA/ALDH2 slow clearance). Since better sleep is neuroprotective for APOE ε4, I upgraded my surgical plans from a minor turbinate procedure (VivAer) to full septoplasty. Mental Health I have, at times, been affected by treatment-resistant depression. Genetics help explain why: I clear stress hormones slowly (MAOA rs6323 TT), have fewer dopamine receptors (DRD2 rs1800497), and 98th percentile anhedonia risk (Ward 2019, Translational Psychiatry, 11 genomic regions). SSRI Underresponder: SSRIs work by boosting serotonin, keeping it active in the brain longer. But my serotonin already clears slowly (MAOA slow), so that's not my bottleneck. My genetics suggest my depression is more about dopamine and the brain's reward system. I have 30-40% fewer dopamine receptors (DRD2), and 98th percentile risk for anhedonia, a reduced ability to feel pleasure. Boosting serotonin doesn't fix a dopamine problem. When Lexapro produced poor results, it wasn't a failure on my part. It was the wrong mechanism for my brain. Ketamine (Genetically Optimal): Ketamine works through a completely different brain pathway than traditional antidepressants. Instead of targeting serotonin or dopamine, it acts on NMDA receptors and glutamate signaling. I also have BDNF Val/Val (rs6265 CC), which means my brain responds well to ketamine-induced neuroplasticity. My genetics confirm IV ketamine is the right treatment, not just a last resort. Prior Anxiety Treatments: Before finding ketamine, I tried other approaches. Guanfacine and Stellate Ganglion Blocks calm the fight-or-flight response. Ativan boosts GABA, the brain's main calming signal. They all helped manage symptoms, but they were treating the downstream effects of anxiety, not the underlying glutamate/dopamine dysfunction that genetics revealed. Lithium Orotate + NAC: My StrateGene report from Sequencing recommended these specifically for my genetic profile. My MAOA is slow at breaking down serotonin, so serotonin lingers too long. Low-dose lithium activates the serotonin transporter (SLC18A1), providing an alternative clearance route that reduces the burden on my slow MAOA enzyme. For dopamine, I have fewer receptors (DRD2). NAC increases dopamine binding to its transporter, which may help my system use dopamine more efficiently. I added 5mg lithium + 200mg NAC daily. Riboflavin (B2): I learned I clear stress hormones like serotonin slower than most (MAOA rs6323 TT). Riboflavin is a helper molecule that the enzyme needs to do its job. I added 36.5mg of the bioactive form (riboflavin 5'-phosphate) daily, though the effect in non-deficient people is unclear. Thyroid: My free T4 is elevated (2.01), with high reverse T3 (34.9), and normal TSH (1.22). Claude helped me understand this isn't hyperthyroidism. It's probably "sick euthyroid syndrome," where the body intentionally converts active thyroid hormone (T4) into an inactive form (reverse T3) during prolonged stress. It's a protective slowdown, not a thyroid problem. Hair Loss I've been thinning since ~2020. Frontal and crown are involved, and it's still progressing despite using a laser cap, peptides, and topical treatments. Genetic testing revealed why some things weren't working and pointed me toward better options. Minoxidil Non-Responder: Minoxidil is a prodrug. It has to be converted by a sulfotransferase enzyme to become active. Testing revealed I lack that enzyme activity (SULT1A1 rs1042028), so the 7% minoxidil in my RootsGA custom compound was doing nothing. I'm upgrading my formula for higher dutasteride and added caffeine, but accepting the wasted minoxidil for now. Shampoo + Conditioner: I was using Act + Acre's Stem Cell System to help with thinning hair, but the formula had weak evidence. I switched to Revita, which has ketoconazole (blocks DHT at the scalp), caffeine, and copper peptides, all with stronger research behind them. Dutasteride Response: Testing predicted 98% response to dutasteride (per RootsGA) based on my SRD5A1/SRD5A2 variants. I'm already using it topically, but this motivated me to upgrade my RootsGA formula from 0.1% to 0.25% dutasteride. Scalp Microneedling: Balding scalp develops fibrosis (scar-like tissue) that can choke hair follicles. I added weekly 0.5mm microneedling, which creates controlled inflammation that remodels this scar tissue and boosts growth factors. I use it before applying my topical for better absorption. CJC/Ipamorelin Bonus: I was already taking CJC-1295/Ipamorelin peptide for body composition. During analysis, Claude pointed out it boosts IGF-1 by 1.5-3x. IGF-1 promotes hair growth and is reduced in balding scalp. This was a hidden benefit I didn't know about. PRP Scheduled: I asked Claude to find therapies beyond my current regimen. It identified PRP (Platelet-Rich Plasma) as well-researched with strong evidence for hair regrowth. I found a clinic in NYC (Great Many) and booked my first session for January. Crohn's & Immunity I've had Crohn's disease since age 10 and I'm on Inflectra (biosimilar infliximab). I get respiratory infections often and frequently have a runny nose. The immunosuppression makes managing infections and inflammation a priority. Theracurmin: I had dropped curcumin from my supplement stack. Claude reminded me why it matters: I have Crohn's and APOE ε4, and curcumin addresses inflammation on both fronts. I re-added it as Theracurmin, a more bioavailable form. PCV21 Vaccine: I'm on Inflectra and get respiratory infections often. Prior vaccines left gaps in my immunity to certain strains (I checked titers). Claude alerted me to a new vaccine (PCV21/Capvaxive) that covers those gaps. I got it. PhytoMulti with Iron: I learned I absorb iron poorly (TMPRSS6 rs855791 AA), explaining why my ferritin has been stuck at 65-72 for years. Crohn's compounds this: gut inflammation impairs iron absorption and chronic inflammation blocks uptake. I was already on PhytoMulti, so I switched to the version with iron (ferrous bis-glycinate, a gut-friendly form). Before → After Brain & Heart • Standard lipid panels → ApoB testing • Weekly calcifediol → D3+K2 3x/week • No brain supplements → CDP choline, glutathione, melatonin-SR • VivAer minor → Full septoplasty (sleep = Alzheimer's prevention) Mental Health • "SSRIs don't work" → "SSRIs don't work because DRD2/MAOA" • Ketamine was working → Confirmed genetically optimal (BDNF Val/Val) • Thyroid looked abnormal → Realized it's stress response • No neurotransmitter support → Lithium 5mg + NAC, riboflavin 36.5mg Hair • RootsGA formula upgraded (dutasteride 0.1% → 0.25%, added caffeine) • Learned 7% minoxidil is wasted (SULT1A1 = can't convert) • Act+Acre → Revita (ketoconazole blocks DHT at scalp) • Only blocking DHT → Microneedling + PRP (targets fibrosis too) • Already on CJC/Ipamorelin → Discovered it has hair benefits Crohn's & Immunity • Dropped curcumin → Theracurmin re-added (anti-inflammatory for Crohn's + APOE ε4) • PhytoMulti → Added iron (genetic + Crohn's malabsorption) • PCV20 → Added PCV21 (covers strains prior vaccines missed)

@DzambhalaHODL Anyone have an MCP connector for DNA data? Or hook up it to a web browser with access to a web interface

Once again, I am pounding the table on using Gemini to analyze your genes. Get a basic Ancestry DNA test, opt into their privacy options, and once you get your results login and download your "raw DNA file" Ask Gemini to give you the identifiers to search for high impact genes and then use it to understand your own data and suggest interventions for the ones with a detrimental impact It's legitimately life changing

@ciphergoth @zynazoid I'm certainly concerned about how to keep up and historically of looked for the hardest challenges I could find.

@zynazoid It does sell itself! But I worry about all my friends who are in denial about what's coming.

Liam, I have been a professional programmer for 36 years. I spent 11 years at Google, where I ended up as a Staff Software Engineer, and now work at Anthropic. I've worked with some incredible people - you might have heard of Jaegeuk Kim or Ted Ts'o - and some ridiculously productive programmers - Eric Biggers, Jeff Sharkey and @jackinwarsaw come to mind as people who seemed to solve problems with code at a truly unearthly rate. At work, I am currently hitting levels of productivity that would put all of them to shame. Not just a rate of making code, but a rate of actually solving problems, that would have been unthinkable two years ago. And it's possible because Claude Code with Opus 4.5 is doing all the heavy lifting; I'm not doing much more than setting direction and reviewing the output. I often have three different sessions going at once, attacking three different aspects of the work I'm doing. Over the holidays I took a break from work Clauding, to do some home Clauding, writing in a few days from scratch a complicated webapp using disparate technologies I had no background in, that would have taken weeks prior to Claude. When I hit problems I just told Claude to debug them and that almost always worked. It also looks great, which is pleasing since not only have I zero CSS skill, I have zero design skill. I'm not out of a job quite yet; there are still some areas where I have better taste than it does, or better instincts. But when you talk about "AI's inability to code", this seems to me to reveal a total disconnect from reality. And this is why I'm urging you to ACTUALLY TRY IT, find out for yourself, and join the rest of us on this Earth.

@ericmigi The only pebble app I wrote, (and still use daily) does exactly this - quick launch, hold down pebble button to voice to text to trello note. But this ring lets me trigger with one hand vs using two.

Pebble Index 01 details - always on you (never need to take it off) - battery lasts for years (doesn't need charging) - open source app - local AI models (no internet connection needed) - customize with actions and MCPs - just $75!

We just launched a new product - @Pebble Index 01 It's a small ring, with a button and mic. Press the button, talk and it saves your thought. I use it many times during the day to jot down reminders, todos, book recs, etc while my hands are full. Just $75 - preorders open today

@ocnj3 @open_erv Big quiet fan

@open_erv BQF?

interesting study on air purifiers in classrooms. They only used enough air purifier capacity to reduce pm pollution by 32% and yet still got 12.5%reduction in sick days (1.2 days per student per year). They used random consumer hepas, so we can do much better. Return on investment in money alone was 9 to 1. According to my crude data/spreadsheet, my BQF based CR box gets 29 times the cubic meters of clean air per dollar than a typical consumer type hepa. So we are up to 261:1 return on investment ratio. What investment is better than that??? Perhaps another public health style thing. We can do both :). Let's do all of the public health things :D. media.rff.org/documents/WP_2…

@Bluntz_Capital robinhood closed their books, bids rejected.

cant have alts go -80% in one candle and not expect a full V reversal after literally every single person got liquidated

@SipeedIO can you tell me when mine will ship? Ordered Dec 3rd...

Happy New Year! 🎉 All #RISCV #NanoKVM products will be in stock in mid-January and are currently on sale at a special price! 🎁 Don't miss out on the perfect New Year's gift: sipeed.com/nanokvm Also, feel free to share your suggestions for NanoKVM Pro! 💬

@TheFandelier Do the different filters all have the same particle filter size? Its hard to tell looking thorugh their site

The filters used in this thread are the Small and Large TruSens Odor & VOC filters. There are frequent sales, on Amazon and the Trusens site. It’s my favorite because it has a truly usable amount of carbon along w hepa a.co/d/h56Dvx0

The Warp Core Family L-R: Warp Core, 46db 100 cfm CADR est Warp Core Maxx, 46db 200 cfm Maxx Core, 40db 70 cfm Micro Core, 37db 30 cfm Micro/Maxx Core breakdowns in thread #daifu

@robwiss @cleanairkits Do you have these spreadsheets somewhere? Esp interested in how thicker filters impact. For example costco has 4/5" 1550s or 1" 2500's.

Ran Filtrete MPR2500, 1900, 1500 and 1085 with a 5-P12 16x25x1 @cleanairkits Luggable. 1900 is CADR king 👑 but 1500 and 1085 are very interesting budget options. 1085 is half the price with a 4-5% CADR loss compared to 1900. 🤔

@TheFandelier Have you seen any data on how different filter thicknesses impact cfm/efficiency for pc fan based boxes?

@open_erv btw, have you seen any data on thickness of filters and their efficiency for PC fan based cr boxes?

@Jolly ok nice, that looks pretty good. No VOC, but it could be part of the solution for sure. I'd like a screen so it can be stand alone though. apolloautomation.com/products/air-1…

Well, I was feeling like a bit of a hypocrite for not actually having a good air quality monitoring device, and was happy to find AirGradient, a supposedly open source option. Then I realized they were not in fact sharing the source code >:(. Any, anywhere.