<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: haldujai</title><link>https://news.ycombinator.com/user?id=haldujai</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Fri, 29 May 2026 17:19:20 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=haldujai" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by haldujai in "Claude Opus 4.8"]]></title><description><![CDATA[
<p>The upper bound is limited by market size and cost of intelligence.<p>Throwing more intelligence at a problem doesn’t necessarily pan out financially otherwise we wouldn’t have single underemployed biology PhD.</p>
]]></description><pubDate>Fri, 29 May 2026 01:47:03 +0000</pubDate><link>https://news.ycombinator.com/item?id=48317974</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=48317974</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48317974</guid></item><item><title><![CDATA[New comment by haldujai in "Claude Opus 4.8"]]></title><description><![CDATA[
<p>You’re putting the cart before the horse - alignment is an unsolved challenge (there are proposed approaches and active research on this) but it is still not established (beyond theory) that latent reasoning is more capable than CoT on hard language reasoning, particularly at scale.</p>
]]></description><pubDate>Fri, 29 May 2026 00:31:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=48317433</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=48317433</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48317433</guid></item><item><title><![CDATA[New comment by haldujai in "Claude Opus 4.8"]]></title><description><![CDATA[
<p>“a bunch of people” aren’t what caused the virus to become less severe.<p>Y2K was overblown how it was portrayed by the media but is irrelevant to the analogy of unsubstantiated extrapolation of early exponential growth.</p>
]]></description><pubDate>Thu, 28 May 2026 21:41:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=48315914</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=48315914</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48315914</guid></item><item><title><![CDATA[New comment by haldujai in "Claude Opus 4.8"]]></title><description><![CDATA[
<p>I think you’re overstating the impact of interpretability here. Your earlier point that latent reasoning models can’t be trained very well and that discretization may be load bearing rather than a readability tax in addition to significant inference infra hurdles (e.g. batching, speculative decoding) have limited any serious attempts and reduced the theoretical advantage over CoT at least in the near term.</p>
]]></description><pubDate>Thu, 28 May 2026 20:38:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=48315127</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=48315127</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48315127</guid></item><item><title><![CDATA[New comment by haldujai in "Claude Opus 4.8"]]></title><description><![CDATA[
<p>I’m also reminded by the early COVID days when exponential growth was leading to predictions of the collapse of modern civilization and a billion dead, now it’s just another endemic respiratory virus.</p>
]]></description><pubDate>Thu, 28 May 2026 20:14:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=48314794</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=48314794</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48314794</guid></item><item><title><![CDATA[New comment by haldujai in "Let's Buy Spirit Air"]]></title><description><![CDATA[
<p>> The current average flight time from NYC to SF is 6.7 hours.<p>What's your source for this? I take this flight a lot and I find it hard to believe it's more than 5.5-5.75 on average. Looking at the last few weeks for one of them[0] supports my experience.<p>Maybe your number has TSA/airport time included.<p>[0] <a href="https://www.flightradar24.com/data/flights/ua1777" rel="nofollow">https://www.flightradar24.com/data/flights/ua1777</a></p>
]]></description><pubDate>Mon, 04 May 2026 07:22:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=48005670</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=48005670</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48005670</guid></item><item><title><![CDATA[New comment by haldujai in "Period tracking app, Flo, found to be selling user data to Meta"]]></title><description><![CDATA[
<p>The bigger gap is for healthcare and business operations which is very broad and includes datasets for AI training as one example.</p>
]]></description><pubDate>Tue, 28 Apr 2026 18:03:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=47938126</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=47938126</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47938126</guid></item><item><title><![CDATA[New comment by haldujai in "Period tracking app, Flo, found to be selling user data to Meta"]]></title><description><![CDATA[
<p>> This one seems clear cut as a HIPAA violation. Glad to hear that interpretation was upheld.<p>Health and wellness apps aren’t covered entities under HIPAA so these disclosures are not violations of it.</p>
]]></description><pubDate>Tue, 28 Apr 2026 16:23:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=47936608</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=47936608</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47936608</guid></item><item><title><![CDATA[New comment by haldujai in "I let ChatGPT analyze a decade of my Apple Watch data, then I called my doctor"]]></title><description><![CDATA[
<p>I wonder if in your case (which is very common) the issue is a mismatch between expectations and reality. The medical system as we know is not designed for someone to listen to you and do a back and forth for hours. If we did that we would only treat 2-4 patients a day. It’s also not particularly helpful.<p>Time spent in a medical encounter is tied to patient satisfaction but there is rapid drop off for clinical benefit especially in the current day where investigations are more important than a physical exam in most cases and more than history in a substantial portion.<p>15 minutes is what we book as follow-ups or minor assessments in US+Canada, usually sufficient for most things. New consults or complex patients are 30-60 minutes.<p>Infodumping is not particularly helpful. Doctors are trained to use a combination of open and closed questions to guide the encounter based on their thinking and understanding of medicine. It’s <i>relevant</i> past medical history as not every symptom or past disease is necessarily useful in assessing what’s wrong today.</p>
]]></description><pubDate>Wed, 28 Jan 2026 18:20:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=46799433</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=46799433</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46799433</guid></item><item><title><![CDATA[New comment by haldujai in "I let ChatGPT analyze a decade of my Apple Watch data, then I called my doctor"]]></title><description><![CDATA[
<p>How so?</p>
]]></description><pubDate>Tue, 27 Jan 2026 02:18:55 +0000</pubDate><link>https://news.ycombinator.com/item?id=46774679</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=46774679</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46774679</guid></item><item><title><![CDATA[New comment by haldujai in "Commonly used arm positions can overestimate blood pressure readings: study"]]></title><description><![CDATA[
<p>> Is the blood pressure really being measured "correctly" in all those studies? Or not?<p>Probably incorrect in most studies, especially large population ones that influence treatment guidelines.<p>It’s academic and doesn’t practically matter though.<p>The pathogenesis of hypertension related disorders (kidney failure, heart failure, stroke etc) is well known.<p>It’s not in doubt that sustained hypertension is bad, that there is increased risk with higher blood pressure and that patients with high blood pressure undergoing treatment suffer less of these bad outcomes.</p>
]]></description><pubDate>Tue, 15 Oct 2024 01:02:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=41843882</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=41843882</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41843882</guid></item><item><title><![CDATA[New comment by haldujai in "Radiology-specific foundation model"]]></title><description><![CDATA[
<p>Yes that’s what I meant too.<p>Acquiring the images is the hard part in obstetrical ultrasound, reporting is very mechanical for the most part and lends itself well to AI.</p>
]]></description><pubDate>Wed, 11 Sep 2024 01:43:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=41507384</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=41507384</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41507384</guid></item><item><title><![CDATA[New comment by haldujai in "Radiology-specific foundation model"]]></title><description><![CDATA[
<p>Potential harm is always the same - misdiagnosis and/or mismanagement.<p>It’s probably very low in the context of CGM and diabetes as the potentially harmful treatments require prescriptions.<p>Device prescription requirements are usually due to product labelling and the manufacturers application. There are OTC fingerstick glucometers and CGMs approved.</p>
]]></description><pubDate>Tue, 10 Sep 2024 14:21:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=41501074</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=41501074</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41501074</guid></item><item><title><![CDATA[New comment by haldujai in "Radiology-specific foundation model"]]></title><description><![CDATA[
<p>> Arguably, prenatal ultrasounds are some of the more challenging to get right<p>Prenatal ultrasounds are one of the most rote and straight forward exams to get right.</p>
]]></description><pubDate>Tue, 10 Sep 2024 13:07:39 +0000</pubDate><link>https://news.ycombinator.com/item?id=41500430</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=41500430</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41500430</guid></item><item><title><![CDATA[New comment by haldujai in "First randomized trial of Ozempic for alcoholism shows big drops in drinking"]]></title><description><![CDATA[
<p>Novo Nordisk is running their lines at max capacity and cutting production of some other drugs. I think the last number I saw was new weekly prescriptions are up 3x from Dec 2023.<p>Unlike Eli Lilly (who use chemical synthesis for Zepbound) they use yeast for synthesis and have decided not to contract out production to protect manufacturing trade secrets which makes ramping up slower, have to wait for their Denmark plant upgrade to come online.</p>
]]></description><pubDate>Mon, 17 Jun 2024 18:47:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=40709480</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=40709480</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40709480</guid></item><item><title><![CDATA[New comment by haldujai in "Auto-brewery syndrome in a 50-year-old woman"]]></title><description><![CDATA[
<p>I’m not sure what you mean by designing diagnostics around an assumed prevalence. The 2017 criteria is pretty broad and I’m not sure how this was influence by assumed prevalence.<p>Is there a reason you suspect a connective tissue disorder to be the unifying diagnosis? I have not heard this theory before.</p>
]]></description><pubDate>Tue, 04 Jun 2024 03:41:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=40570464</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=40570464</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40570464</guid></item><item><title><![CDATA[New comment by haldujai in "Auto-brewery syndrome in a 50-year-old woman"]]></title><description><![CDATA[
<p>> I think a large number of those with undiagnosed conditions actually exhibit conditions from the hEDS bucket of comorbidities which is especially vast and varied. See <a href="https://ohtwist.com/about-eds/comorbidities" rel="nofollow">https://ohtwist.com/about-eds/comorbidities</a> for an incomplete list. This bucket does include SIBO. So if you have been told by a doctor that you're a hypochondriac and you exhibit SIBO these do increase the probability of you having hEDS.<p>Not really. SIBO prevalence increases with age and goes up to 80% in elderly patients.<p>hEDS prevalence is between 0.005% and 0.02%.<p>The symptoms/diseases you listed are very common. We don’t really use hypochondriac anymore but the reason these are common is because they’re vague constitutional symptoms that have 1000 different possible causes or nothing at all.<p>> What is very unorthodox about about my view is that this flexibility is not an essential component so a lack of flexibility is not sufficient to rule it out even if it does reduce the likelihood.<p>There is nothing unorthodox about this, it’s part of the 2017 international criteria for hEDS.<p><a href="https://www.ehlers-danlos.com/wp-content/uploads/2022/03/Malfait_et_al-2017-American_Journal_of_Medical_Genetics_Part_C__Seminars_in_Medical_Genetics.pdf" rel="nofollow">https://www.ehlers-danlos.com/wp-content/uploads/2022/03/Mal...</a></p>
]]></description><pubDate>Mon, 03 Jun 2024 15:19:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=40563529</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=40563529</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40563529</guid></item><item><title><![CDATA[New comment by haldujai in "mRNA Cancer Vaccine Reprograms Immune System to Tackle Glioblastoma in 48 Hours"]]></title><description><![CDATA[
<p>You’re missing that we have treatments with known risk/benefits vs research treatment X with undefined risk and reward.<p>It’s extremely challenging if not impossible to obtain informed consent in this situation.<p>We’ve been through this before where a fancy new treatment with promising early/lab results usurped conventional therapy only to later be found inferior.<p>Earlier TKIs and NSCLC are a recent example that comes to mind.</p>
]]></description><pubDate>Thu, 02 May 2024 10:15:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=40234518</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=40234518</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40234518</guid></item><item><title><![CDATA[New comment by haldujai in "FTC challenges 'junk' patents held by 10 drugmakers, incl Novo Nordisk's Ozempic"]]></title><description><![CDATA[
<p>It’s similar in both countries with some slight variation in the US depending on the state.<p>Unless the prescription is marked as “do not substitute” pharmacists generally have the discretion to substitute for a generic, in fact a few states require it.</p>
]]></description><pubDate>Tue, 30 Apr 2024 22:34:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=40217248</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=40217248</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40217248</guid></item><item><title><![CDATA[New comment by haldujai in "Cost of developing new drugs may be lower than industry claims: trial"]]></title><description><![CDATA[
<p>Sexual dysfunction is a <i>real, widespread disease</i>.<p>With that said, Viagra was initially developed for high blood pressure and chest pain. Testing didn’t pan out but it worked wonders for erections so it was rebranded. It is also used in pulmonary hypertension.</p>
]]></description><pubDate>Sun, 28 Apr 2024 09:58:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=40187378</link><dc:creator>haldujai</dc:creator><comments>https://news.ycombinator.com/item?id=40187378</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40187378</guid></item></channel></rss>