HNHacker News
TopNewBestAskShowJobs

faangguyindia

912 karma · joined July 3, 2022

email: coredrop@proton.me
submissionscomments
faangguyindia··on Biggest Probabilistic Computer Turns Noise into Answers
curious, where the guys who are working on mathematical models run their simulations these days.
faangguyindia··on Popular sugar substitutes linked to faster brain aging
I eat fruits but i also eat high quality white cane sugar.
faangguyindia··on Codex Resets
yes but as a normal user your data and traces are being used to train AI which helps enterprise customers.
faangguyindia··on Popular sugar substitutes linked to faster brain aging
I've never relied on sugar free foods, and I consume plenty of natural sugars and easily cut weight on a regular basis.

And often people tell me not to eat "suger" and carbs like "rice". I never listen to them.

People find it hard to lose weight and it's not hard to understand why.

Nothing beats a calorie deficit for weight loss.

Deficit = TDEE (total calorie burn for the day) - calorie intake

Most people go wrong when they diet. These are the reasons I see most often:

a) They try to eat too little. It works in the short term, but within several weeks, they become miserable and bounce back to binge eating, undoing all their progress.

Many people have a TDEE of 2,700 kcal and start eating an 800–1,100 kcal diet for weight loss.

Eating 800–1,500 kcal isn't wrong, but it's more suitable for someone who is an experienced dieter. Beginners, more often than not, will fail with this approach.

b) They don't know the calorie density of foods. Many believe granola is a low calorie, healthy option based purely on their gut feelings about the food and its appearance.

c) Many people are sedentary and have low muscle mass, so their TDEE is quite low. To cut weight, they need to go 400–500 kcal below it, and guess what? They end up miserable because their TDEE is already low, and lowering their intake by 400–500 kcal puts them in a zone with too little food.

The solution is Zone 2 cardio not because it burns more fat, but because it's easy on the joints, and most people can sustain it for hours, resulting in a higher cumulative calorie burn with much less stress on the recovery system.

To throw a monkey wrench into this scheme, oftentimes when you diet down, your body lowers its metabolism (through metabolic adaptation or increased efficiency in physical activities). Other times, people simply move less when they eat at a deficit because they feel low on energy, lowering their TDEE.

Remember, since deficit = TDEE - intake:

If TDEE drops too low, calorie intake must go down to maintain the deficit; otherwise, there will be no weight loss.

I've been working to solve this problem and we've lots of success with maintenance tracking approach.

faangguyindia··on Popular sugar substitutes linked to faster brain aging
isn't that easy to fix by simply comparing them to a group who are also overweight and cut calories while still having no artificial suger.
faangguyindia··on What AI did to stackoverflow in a graph
It’s very easy to destroy a specific sub (if it’s not too big and doesn’t have half a dozen mods).

it's like those ddos rings, but works on social networks.

You can create a sub and a Discord group, then ask people in the Discord group to launch a mass report against your competing sub and its moderators. You can use scanners to find questionable stuff that you can report, and more often than not, this will get the mod banned. If the sub doesn’t have multiple mods (with unique IPs, as Reddit tracks fingerprints), the sub is now in the hands of Reddit’s mod team.

Back then it was not this easy, but now with AI and residential IPs you can create lots of fake users and reports etc... and take almost any avg redditor down.

faangguyindia··on What AI did to stackoverflow in a graph
often people make a sub, it gets popular because of specific content. then the original mod gets banned, now their sub is orphan and reddit team assigns it to someone else.

many times you'll notice the new mod became active as a contributor on specific sub only 2-3 days before the OG mod gets banned and sub declared orphan.

Coincidence?

How can reddit even hand over a sub to someone who has nothing to do with a community? Simply because they install and control who controls the sub.

Mods get "comment/post removal" power, so they use it to shape the community towards specific "narrative", there is no audit trail for any mod specific actions unless you are a mod you perhaps can't see what all a mod is doing on a sub.

Also, they can simply make an automod/bot rule which simply removes your comment by creating a rule with your username after that you'll not know your comment is gone but others will not see it!

faangguyindia··on Ask HN: What are you working on(2026)?
yes, daily weighing helps a lot more so initially.
faangguyindia··on What AI did to stackoverflow in a graph
Reddit is on the same track. What I've noticed is that moderators have become increasingly hostile. Reddit's AI moderation, which is designed to remove AI slop, has removed multiple top contributors I used to follow on programming, electronics, bodybuilding, welding, machining, and 3Dprinting subs. And the worst part is that Reddit site admins have no idea about this. I know a guy who reckons he can get any account banned from Reddit by simply mass reporting it using residential IP providers and OpenClaw.

Good luck trying to write any helpful posts in the community anymore, someone will come along and respond with "AI."

faangguyindia··on Ask HN: What Are You Working On? (July 2026)
Guaranteed weight loss (for overweight) and weight gain for (underweight) people, you'll see clear gain or loss within 2-5 weeks.

If you are in a calorie surplus, your weight goes up. If you are in a calorie deficit, your weight goes down. MacroCodex uses that relationship to estimate your true maintenance calories over time: https://macrocodex.app/ This can be used to gain or lose weight. You simply need to eat above your maintenance calorie line to gain weight and below it to lose weight. No ads, no subscriptions, no payments, 16k users already.

If you're interested in how it works: https://macrocodex.app/knowledge/rethink/adaptive-tdee/

faangguyindia··on Ask HN: What are you working on(2026)?
Usually in 16-21 days
faangguyindia··on Ask HN: What are you working on(2026)?
If you are in a calorie surplus, your weight goes up. If you are in a calorie deficit, your weight goes down. MacroCodex uses that relationship to estimate your true maintenance calories over time: https://macrocodex.app/

This can be used to gain or lose weight. You simply need to eat above your maintenance calorie line to gain weight and below it to lose weight. No ads, no subscriptions, no payments, 16k users already.

If you're interested in how it works: https://macrocodex.app/knowledge/rethink/adaptive-tdee/

faangguyindia··on SQLite Is All You Need
lobster.rs also started using sqlite, i've been using exclusively using sqlite for most of my projects.

apps appear much faster because db is on the same server

faangguyindia··on The privacy problems hidden in your period tracker
>I do find that the best apps are the paid ones, they don't need to be relying on selling user data to make a profit.

There is no limit to greed, you can find billionaires who are after trillions. This line of reasoning is at fault.

faangguyindia··on The privacy problems hidden in your period tracker
>Yes, backing up is the most obvious reason to have it live on a server from the user POV. And of course the app vendor will want more lockin and be able to keep the data hostage.

Android comes with auto backup, https://developer.android.com/identity/data/autobackup

App developers choose to collect more data than they need.

faangguyindia··on The privacy problems hidden in your period tracker
You can collect data without collecting information about who the data belongs to if you only need it to improve your statistical model. That's what we do. We do not know the user's name, phone number, email address, or date of birth. I think it's best to not collect stuff you don't need for your app to work.
faangguyindia··on The privacy problems hidden in your period tracker
Most people won't pay $70 a year for an app which uses Gemini API to track calories.

That's something you can directly do in AI studio for free, or use free API quota.

faangguyindia··on An Engineer's Guide to USB Typе-С (2024)
turns out it was just me not knowing my keyboard has full qmk support and can be configured via usevia, what N-key rollover (NKRO) is.

my keyboard is actually r65 by royal kludge i got that wrong!

I was gifted this keyboard and i just love its feel.

faangguyindia··on An Engineer's Guide to USB Typе-С (2024)
You are genius! thanks this completely solved my issue, for those curious https://postimg.cc/75hV9tCK

I use usevia, all i had to do was find a keyboard definition file for my keyboard which claude found in 2 minutes!

faangguyindia··on An Engineer's Guide to USB Typе-С (2024)
I use an Apple Silicon Mac and often use programmable keyboards like the Royal Kludge RK61 via USB C. when I press keys such as A, S, D, F, W, or nearby keys in quick succession, the keyboard stops responding completely until I unplug and reconnect it. I've even replaced the USB C cable with a new store bought cable, but the issue still persists.
faangguyindia··on Solod: Go can be a better C
i've been using Go on backend now all the time.

Wish something like this existed instead i am stuck with flutter and react native on Mobile.

When will a time come when i can use some functional languages like Haskell or a plain boring language like Go for making apps with OTA ability for mobiles.

As vibe coding takes over, app store approval will become slowerer and OTA is really great when you need to make quick changes!

You can OTA each day and do base app release to store once each week.

i think this maybe the space where very little work is being done.

faangguyindia··on Google Kills Custom Search API on Jan 1, 2027
I was using it for an ai agent i built for my use as home assistant. It seems like many competing companies might be doing same.
faangguyindia··on My car’s OTA update broke Android Auto
>That said shipping JIT / interpreter with your program to recompile updates / parts of it sounds silly to me.

why is it silly? modern phones can easily do this (they are running local llm, what issue is with jit?), if it's done one time per update what's the problem?

what's the problem with jitting couple of hot paths which have been updated and were not part of base app?

>Apple does not allow running code from internet either (even if interpreted iirc)

And how is React native expo/eas updates are working then? Isn't it downloading a bytecode bundle?

Afaik, interpreter is allowed. JIT isn't?

faangguyindia··on My car’s OTA update broke Android Auto
no, i am not asking for AOT OTA update.

AOT will be in base app and it will include JIT or Interpreted OTA updates.

For Apple, JIT can simply be disabled and OTA update can run patched part in interpreter.

But JIT works on Android (well), so this Hybrid system is capable of being much faster than React Native where your JS code only runs in Hermes VM which isn't JIT.

In this system, all your crazy math and algorithms (on hot path) stay easily updatable as this small part can be run on JIT with snapshot saved for next load if it has not changed!

JIT on average is 100x faster than dumb interpreter for language like Dart.

Does it already exist? otherwise i am seriously thinking about building something like this.

faangguyindia··on My car’s OTA update broke Android Auto
why OTA update OS that frequently?

I've been lately into mobile apps and i am finding that there is no system which combines these 3

1. AOT 2. JIT (for hot paths) 3. Interpreter for non JIT paths or where you explicitly do not want jit.

Imagine, a system which compiles your app to AOT but when you push OTA update, part of the app are selectively replaced to JIT or Interpreted mode.

it's theoretically possible but nobody seems to be doing it. I found react native / expo eas update but i don't think it's like this, it has a Hermes VM which runs bytecode but it has no JIT so you'll write native code for hot path then you'll need to upload a full update to Android. So, only toy level code performance can be can actually be written in JS?

Much better, patch the parts where AOT calls into JIT or interpreter.

Currently i am using react native and flutter. Flutter's UI framework code is in Dart if you load this whole code into JIT, it will consume a lot of resources on mobile device as the framework is big and does lot of work. If all framework code was AOT and your custom patchable code also comes with AOT but upon OTA replaced by JIT or Interpreted code, crazy performance!

But what if we could run the most of the code in AOT and only run changed code in JIT or interpreted mode? arguably it would perform as good as it does not being complete AOT while also providing react native like fast updates.

faangguyindia··on The privacy problems hidden in your period tracker
Greg isn't the only person who owns MF, there are people like Nippard who call it (my app and are either shareholder or promoter with a lot of influence), if you look at the whole hussein and julian fiasco and solomon nelson episode well...

This is kinda misleading. First of all, MacroFactor uses the Gemini API for photo to calorie/macro estimation, so the data does leave their premises. It does not work completely offline, so it does call home, and after that, you never really know what happens with the data.

MacroFactor charges roughly $70 a year. MacroCodex is free and doesn't require an internet connection to work. It can also offset random weight gain due to PMS or other short term hormonal cyclical issues, as well as other water retention issues. It doesn't even ask for your phone number or email or even date of birth!! (it just asks age) on Android. On the web app, an email is required only for storing your data (due to the volatile web storage offered in PWAs, where the OS/browser can evict storage under memory pressure).

If an app doesn't collect your personal identifiable data it cannot sell it, if it's capable of running offline it can be put behind a firewall rule and/or diagnostic/telemetry data disabled in setting though i'd argue if it's not collecting your personal info and want data for improvement of specific app feature (which benefit from data analysis) then you should perhaps analyze the risk of this decision.

faangguyindia··on Your 'app' could have been a webpage (so I fixed it for you)
yes, i am forced to to make a real app because storage is not reliable in PWA, browser or OS can wipe off data.

i don't want to pay for servers just to have an app.

and updating apps is slow, for flutter you need to pay for shorebird.

In react native land, not sure but there are paid stuff like expo? you can self host but usually you end up payign for some OTA provider?

faangguyindia··on What xAI's Grok build CLI sends to xAI: A wire-level analysis
this might be relevant about microsoft.

Federal Cyber Experts Thought Microsoft’s Cloud Was “a Pile of Shit.” They Approved It Anyway:

https://www.propublica.org/article/microsoft-cloud-fedramp-c...

faangguyindia··on What xAI's Grok Build CLI Actually Sends to xAI
How did book publishers figure out ai stole from them? Can people use a similar way to figure out if their private repos have become part of the training corpus?
faangguyindia··on What xAI's Grok build CLI sends to xAI: A wire-level analysis
Your trade secret is already gone the moment you unleashed non local ai agents on your codebase.

This is why I keep a separate repo for important parts that I do not want competitors to get access to, and only use ai on dumb parts which I don't care if get leaked tomorrow.

← PreviousPage 5 of 27Next →