Siddhartha retweeted
>AI causing RAM shortages
>Devs start giving a fuck about memory optimization again
>AI starts coding extremely efficient apps across all platforms
This is accelerationism in action
Point 3.
Hello everyone. Happy Diwali!
Thank you for your criticism, your support and your time over the last two days. Gives me much needed clarity on what we need to explain better:
1. Temple is not made for India. It’s made for the world. Please check pricing. We don’t expect any more than 1% of Temples being sold in India.
2. If you are upset that the pricing is too high, I am sorry. There’s a lot of work that went into Temple, and we priced it at what we think it’s worth. If you think its overpriced, you can wait for user reviews to decide if its worth your money.
3. Please do not support us because we are an Indian startup. We want to compete globally. We don’t need/want nationalist support. Support us if you want to support early innovation, which could help you live a better life, and will continue to add more proof to its science over the years.
4. There are different audiences which buy products at various levels of proof. Proof goes from whitepaper -> pre-print -> peer reviewed. You are free to sit on the fence until peer reviewed proof arrives. You are free to not believe it even then. Nobody is forcing you to buy. Making announcements is our right; and if we overclaim anywhere, we will inevitably fall into regulatory trouble everywhere in the world.
5. We didn’t make any medical claim. Doctors primarily see things from a medical lens, so they see every claim as a medical claim. I repeat – Temple is strictly a fitness + AI device, and we've made no medical claims. Temple is not proven to diagnose or prevent any disease. It doesn't do anything which replaces the need for you to see a doctor.
6. The main point of Temple is NOT Brain Flow. That’s what we just started Temple for, and studies on Brain Flow are the first ones getting out.
7. The main point of Temple is Entropy (I shared the HomeScreen of Temple three months ago with Entropy as the main thing). Our emails to our waitlisted users only talk about Entropy. We haven’t published anything on Entropy yet. We will. If you want to buy Temple before that, please do. If you don’t want to, don’t.
8. Roomdeaf AI is an expansion of the device’s capabilities, and our vision for Temple. It is not a pivot.
9. Social media is not the regulator for Temple. Wherever we need regulatory clearances, we will have them before we make any claim needing them.
10. I will post a detailed note on how we calculate Brain Flow, its uses, and most importantly, its limitations in a bit (tomorrow).
Siddhartha retweeted
Me acaba de volar la cabeza esto…
Apple didn't store a plaintext list of forbidden words in iOS. Instead, researchers discovered a compiled minimal perfect hash table of 32-bit fingerprints buried inside the keyboard's language-model files.
When you swipe-type or use autocomplete on an iPhone, the keyboard converts your input into token IDs and evaluates sequences of two to six words. It hashes those candidate n-grams using Bob Jenkins's public-domain hashword() function. If that 32-bit fingerprint matches an entry in the table, the predictive engine refuses to suggest it, or the swipe model actively diverts the word to an alternate match.
On November 6, researcher Derik Kauffman published Swipeblocked after exhaustively enumerating candidate word combinations against the table to reverse-engineer the hidden phrases. His conservative estimate is roughly 100,000 entries across 28 languages. In English, the prediction list holds 4,525 phrases and the swipe-typing list holds 2,794.
Most entries cover violence, slurs, and sexual content, but the table also silences geopolitical phrases like "Tibetan independence", "United Ireland", "dictator Trump", and Hebrew for "the occupied territories". In testing, swiping "United Ireland" consistently resolved to "United Iceland".
You can still type any phrase letter by letter. The suppression only touches autocomplete predictions and gesture paths. Apple told Reuters the system avoids offensive suggestions, but never addressed the political entries or the existence of the table.
Apple never published the list because it was never meant to be read. It took brute-forcing the hash table to show what the keyboard was quietly dropping.
Siddhartha retweeted
Comparing US AQI standard (left) vs NAQI (National Air Quality Index India).
Same pollutants, different representation. We’re also switching to US AQI standard as default as per requests.
Siddhartha retweeted
thor love and thunder is like everyone involved took it as a joke except for christian bale who put his soul into it
Apple uses map data from mapmyindia for India fyi
Siddhartha retweeted
No one questioned Whoop or Ultrahuman, but are dunking on Temple even before its full release.
Whoop did not have a clinical validation study for 5 years after the release. Ultrahuman still doesn't have one.
These are just trackers, not diagnostic devices, used by a bunch of niche enthusiasts to improve habits, not treat medical issues.
If they scale to become more widespread and plan to get closer to diagnosis (like Apple or Whoop adding ECG), then it makes sense to do intense peer-reviewed clinical studies.
Let folks experiment and innovate. Basic checks and balances have to be there, of course. But you cannot build such a high barrier to entry. Then we won't see any progress.
We love to pull each other down. A shitty culture ingrained in us. We need to do better. Encourage, in a balanced way. No blind worship but constructive feedback. Not call everything a scam at the drop of a hat, but patiently understand what is trying to be achieved give a mature, reasoned take on what you agree with, what you don't agree with, what you don't know but will wait and watch. This kind of dialogue is missing. Sad.
Siddhartha retweeted
One guy put $20M of his own money into building something in India to push science forward.
On the other side: a comedian whose biggest product is outrage, and a doctor famous for being angry online. They've already ruled on a device that hasn't shipped.
Every country that leads in tech and science had founders called frauds before the data came in. Some were wrong. Some changed the world. Nobody knew until they were allowed to try.
India keeps asking why we don't build our own Apple, our own Neuralink. This is why.
The moment someone tries, the timeline becomes a courtroom, and the judges have never built anything that had to work.
Ask for data? Absolutely. But "scam" before launch isn't scrutiny. It's content.
In a few years, Temple may be proven right or wrong. Either way, one man will have built something and maybe moved forward how we understand ageing.
The other two will still be quote-tweeting the next person who dares to try.
Pick your side. I've picked mine.
Siddhartha retweeted
- Restaurants have no moat.
- Anyone can vibecook their own meal.
- You can build your own burger at home at a fraction of the cost it takes you to get it from a restaurant.
- If a restaurant builds a really good recipe, copycats will instantly reverse taste it and steal it.
Restaurants are dead.







































