Q: Why are we building this?
A: To design better, safer pesticides. The therapeutics industry has seen huge improvements by integrating machine learning models into its drug discovery pipelines; on the other hand, the agrochemical pesticide discovery space has seen almost no improvement, despite the high degree of similarity between the two processes. This is because the agrochemical industry is largely stagnant; fewer than 40 new pesticide ingredients have been developed in the last decade, with most "new" pesticides being just tweaks of existing ones. To solve this, we're using our SOTA machine learning models to find radically better pesticides.
Because this is a search problem, speed is key. This is why we've spent so much time optimizing our models to be as fast as possible, which also allows us to make them freely available without the long queue times or rate limits alternatives like AlphaFold 3 or Chai have.
Q: Who are we?
A: We\u2019re Navvye Anand (Caltech) and Tyler Rose (Wolfram Research), a scrappy duo of engineers who met at the Wolfram Summer Research Program in June 2023. We\u2019re Indo-Chinese founders who both grew up visiting farmlands in our countries. United by our passion for tackling hard problems, we dropped out of college and started Bindwell to transform the archaic agrochemical industry.
We\u2019re joined by Max, a prodigious hacker and open source contributor who dropped out of a math degree at Reed College. Max is relentlessly curious about everything from robotics to philosophy, and like us, his passion for solving challenging problems made him a perfect fit for the team.
We'd love to hear from you, whether it's feedback about our app, questions about our mission, or anything else you think is interesting. You can reach us by commenting on this post, or by emailing us at founders@bindwell.ai."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Interactive Drug Screening in Your Browser"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://app.bindwell.ai/"}},"_tags":["story","author_niederman","story_43177281","show_hn"],"author":"niederman","children":[43179945],"created_at":"2025-02-25T20:55:28Z","created_at_i":1740516928,"num_comments":2,"objectID":"43177281","points":3,"story_id":43177281,"story_text":"Today, the team at Bindwell (YC W25) is launching a free, unified platform for finding candidate drugs (or pesticides, or herbicides, or anything else that needs to bind to a protein), and visualizing their structure. Unlike AlphaFold 3 and others, we've worked hard to make the experience as interactive as possible by designing and training our own accelerated models, significantly speeding up the candidate screening workflow.
Q: Why are we building this?
A: To design better, safer pesticides. The therapeutics industry has seen huge improvements by integrating machine learning models into its drug discovery pipelines; on the other hand, the agrochemical pesticide discovery space has seen almost no improvement, despite the high degree of similarity between the two processes. This is because the agrochemical industry is largely stagnant; fewer than 40 new pesticide ingredients have been developed in the last decade, with most "new" pesticides being just tweaks of existing ones. To solve this, we're using our SOTA machine learning models to find radically better pesticides.
Because this is a search problem, speed is key. This is why we've spent so much time optimizing our models to be as fast as possible, which also allows us to make them freely available without the long queue times or rate limits alternatives like AlphaFold 3 or Chai have.
Q: Who are we?
A: We\u2019re Navvye Anand (Caltech) and Tyler Rose (Wolfram Research), a scrappy duo of engineers who met at the Wolfram Summer Research Program in June 2023. We\u2019re Indo-Chinese founders who both grew up visiting farmlands in our countries. United by our passion for tackling hard problems, we dropped out of college and started Bindwell to transform the archaic agrochemical industry.
We\u2019re joined by Max, a prodigious hacker and open source contributor who dropped out of a math degree at Reed College. Max is relentlessly curious about everything from robotics to philosophy, and like us, his passion for solving challenging problems made him a perfect fit for the team.
We'd love to hear from you, whether it's feedback about our app, questions about our mission, or anything else you think is interesting. You can reach us by commenting on this post, or by emailing us at founders@bindwell.ai.","title":"Show HN: Interactive Drug Screening in Your Browser","updated_at":"2025-02-26T03:26:51Z","url":"https://app.bindwell.ai/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"fractal618"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["chai","discovery"],"value":"So. Sometimes we have to admit defeat and move on, roll with the punches and even admit/accept defeat. But I will never forget getting library lunch passes from my homeroom teacher so I could play illegally installed Diablo 1 and age or empires 1 on the library terminals with a few other wonderful nerdy 8th grade colleges. The times of shared phone lines and AOL dialup chat rooms, instant messaging and ARC (attack retrieve capture). To scroll through geocities pages and discover the very indie game Silencer <3. If I could go back I most certainly would. Not only to bring the hard earned lessons of dating and relationships. Or the stock market predictions. But just to be naive and hopeful again. Kazaa! ROFL. I\u2019m not a bad\nGuy. Just delayed and lost and missed so many opportunities for all the RIGHT reasons. I remember the innocence and justification of being a young digital explorer and I\u2019m fucking grateful for that. No regrats not even one letter!"},"title":{"matchLevel":"none","matchedWords":[],"value":"What I Can Remember"}},"_tags":["story","author_fractal618","story_49762163","ask_hn"],"author":"fractal618","children":[49762315,49762317,49768717],"created_at":"2026-09-19T00:40:09Z","created_at_i":1789778409,"num_comments":3,"objectID":"49762163","points":8,"story_id":49762163,"story_text":"So. Sometimes we have to admit defeat and move on, roll with the punches and even admit/accept defeat. But I will never forget getting library lunch passes from my homeroom teacher so I could play illegally installed Diablo 1 and age or empires 1 on the library terminals with a few other wonderful nerdy 8th grade colleges. The times of shared phone lines and AOL dialup chat rooms, instant messaging and ARC (attack retrieve capture). To scroll through geocities pages and discover the very indie game Silencer <3. If I could go back I most certainly would. Not only to bring the hard earned lessons of dating and relationships. Or the stock market predictions. But just to be naive and hopeful again. Kazaa! ROFL. I\u2019m not a bad\nGuy. Just delayed and lost and missed so many opportunities for all the RIGHT reasons. I remember the innocence and justification of being a young digital explorer and I\u2019m fucking grateful for that. No regrats not even one letter!","title":"What I Can Remember","updated_at":"2026-09-20T08:35:10Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jwally"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["chai","discovery"],"value":"I'm feeling pretty depressed right now, and hopefully sharing this little anomaly I discovered in January might help someone make the world a better place, and figure out how to put a dent in online crime. Fuck criminals.
Most residential proxy providers that I've seen don't support UDP. UDP is what WebRTC uses and has been part of the browser's "core" since 2021. If someone says they're on Chrom(), FireFox, Safari - they have it.
So the setup goes like this:
1) Create a subdomain to get the client's TCP ip address. You can force it to ipv4 by not giving it an AAAA record. This will make future-comparison easier.
2) Create a subdomain to get the client's UDP ip address. Ditto the quad-a bit. You'll want to run your own STUN server (are you there, Claude? Its me, Justin...). To make it tamper resistant, I have the STUN server encrypt the ipv4 address and some other stuff and return it as an ipv6 ice candidate. Its abuse, but it works and I don't have to worry about back-end-server communication bottle necking anything. Just catch, encrypt, return.
And the algorithm looks like this:
** YOU HAVE A TCP AND UDP IP ADDRESS ***
1) If both IP Addresses are the same - probably not a residential proxy
2) If both IP Addresses are different but on the same ASN and in the same .../24(?) - they're on a cell network
3) If the IP Addresses are on different ASNs - that's a MAJOR red flag. Most people who use residential proxies have it drilled into their head - TURN OFF WebRTC. This leaked bad in the mid 2010's and people are still terrified of it.
4) If one of the Addresses is on Cisco Web Umbrella or zScaler - and other is residential - they're behind a corporate protection firewall. Not really a proxy per se, but it'll trip this algorithm. Based on your threat model, they're probably ok to whitelist.
*** YOU __ONLY__ HAVE TCP IP ***
So this is the "one weird trick" / secret sauce / whatever that makes this work. TLS-duration / TCP-rtt > 3.0
* "Many residential proxies split the connection at the exit: the origin\u2019s TCP socket ends there, so Linux TCP_INFO estimates only origin <=> exit RTT. The exit forwards TLS handshake bytes to the true client, making ClientHello handshake-complete include the hidden proxy chain. An unusually high TLS-duration/TCP-RTT ratio exposes that split." *
This is really all we need. From here, we just do the following:
TCP == UDP === not proxy
TCP ~= UDP === if on same ASN, not proxy
TCP != UDP === proxy or zScaler-thing. Whitelist to suit.
No UDP and TLS/TCP RTT Ratio < 3 === probably not a proxy (lower is better here)
No UDP and TLS/TCP RTT > 3 === probably a proxy (most I've seen float between 6-12x)
I've worked with this thing for the last 6 months and have worked mainly with Soax, BrightData, and OxyLabs proxies. I also built my own ProxyDevice (android) and C2 - put it on my own network and it caught it.
This is extremely cool because this means you don't have to blacklist IP addresses (not like it works) since a legitimate user and a proxy will come from the same IP - the proxy client will look _very_ different. They get blocked by NetAcuity/MaxMind/StaticList - the entire IP is killed. This is surgical.
This probably isn't perfect, but based on my limited experience its spooky accurate. I won't claim its 100% accurate for all cases everywhere but if its better than nothing - by definition its better than nothing - and that helps. Which, at the end of the day is all I want to do with this - help.
p.s. more than happy to share what I've built. Test, evidence, code, whatever. Just let me know."},"title":{"matchLevel":"none","matchedWords":[],"value":"How to Detect Residential Proxies"}},"_tags":["story","author_jwally","story_49668557","ask_hn"],"author":"jwally","children":[49672239,49693563,49693818],"created_at":"2026-09-12T03:45:23Z","created_at_i":1789184723,"num_comments":2,"objectID":"49668557","points":4,"story_id":49668557,"story_text":"I'm feeling pretty depressed right now, and hopefully sharing this little anomaly I discovered in January might help someone make the world a better place, and figure out how to put a dent in online crime. Fuck criminals.
Most residential proxy providers that I've seen don't support UDP. UDP is what WebRTC uses and has been part of the browser's "core" since 2021. If someone says they're on Chrom(), FireFox, Safari - they have it.
So the setup goes like this:
1) Create a subdomain to get the client's TCP ip address. You can force it to ipv4 by not giving it an AAAA record. This will make future-comparison easier.
2) Create a subdomain to get the client's UDP ip address. Ditto the quad-a bit. You'll want to run your own STUN server (are you there, Claude? Its me, Justin...). To make it tamper resistant, I have the STUN server encrypt the ipv4 address and some other stuff and return it as an ipv6 ice candidate. Its abuse, but it works and I don't have to worry about back-end-server communication bottle necking anything. Just catch, encrypt, return.
And the algorithm looks like this:
** YOU HAVE A TCP AND UDP IP ADDRESS ***
1) If both IP Addresses are the same - probably not a residential proxy
2) If both IP Addresses are different but on the same ASN and in the same .../24(?) - they're on a cell network
3) If the IP Addresses are on different ASNs - that's a MAJOR red flag. Most people who use residential proxies have it drilled into their head - TURN OFF WebRTC. This leaked bad in the mid 2010's and people are still terrified of it.
4) If one of the Addresses is on Cisco Web Umbrella or zScaler - and other is residential - they're behind a corporate protection firewall. Not really a proxy per se, but it'll trip this algorithm. Based on your threat model, they're probably ok to whitelist.
*** YOU __ONLY__ HAVE TCP IP ***
So this is the "one weird trick" / secret sauce / whatever that makes this work. TLS-duration / TCP-rtt > 3.0
* "Many residential proxies split the connection at the exit: the origin\u2019s TCP socket ends there, so Linux TCP_INFO estimates only origin <=> exit RTT. The exit forwards TLS handshake bytes to the true client, making ClientHello handshake-complete include the hidden proxy chain. An unusually high TLS-duration/TCP-RTT ratio exposes that split." *
This is really all we need. From here, we just do the following:
TCP == UDP === not proxy
TCP ~= UDP === if on same ASN, not proxy
TCP != UDP === proxy or zScaler-thing. Whitelist to suit.
No UDP and TLS/TCP RTT Ratio < 3 === probably not a proxy (lower is better here)
No UDP and TLS/TCP RTT > 3 === probably a proxy (most I've seen float between 6-12x)
I've worked with this thing for the last 6 months and have worked mainly with Soax, BrightData, and OxyLabs proxies. I also built my own ProxyDevice (android) and C2 - put it on my own network and it caught it.
This is extremely cool because this means you don't have to blacklist IP addresses (not like it works) since a legitimate user and a proxy will come from the same IP - the proxy client will look _very_ different. They get blocked by NetAcuity/MaxMind/StaticList - the entire IP is killed. This is surgical.
This probably isn't perfect, but based on my limited experience its spooky accurate. I won't claim its 100% accurate for all cases everywhere but if its better than nothing - by definition its better than nothing - and that helps. Which, at the end of the day is all I want to do with this - help.
p.s. more than happy to share what I've built. Test, evidence, code, whatever. Just let me know.","title":"How to Detect Residential Proxies","updated_at":"2026-09-20T13:18:10Z"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["chai"],"value":"anie_cha"},"title":{"matchLevel":"none","matchedWords":[],"value":"New Function of Blinking Identified"},"url":{"fullyHighlighted":false,"matchLevel":"partial","matchedWords":["discovery"],"value":"https://www.psypost.org/scientists-have-discovered-a-previously-unknown-function-of-blinking/"}},"_tags":["story","author_anie_cha","story_40774217"],"author":"anie_cha","children":[40774218],"created_at":"2024-06-24T10:14:41Z","created_at_i":1719224081,"num_comments":0,"objectID":"40774217","points":2,"story_id":40774217,"title":"New Function of Blinking Identified","updated_at":"2024-09-20T17:19:20Z","url":"https://www.psypost.org/scientists-have-discovered-a-previously-unknown-function-of-blinking/"}],"hitsPerPage":50,"nbHits":16,"nbPages":1,"page":0,"params":"query=Chai+Discovery&tags=story&hitsPerPage=50&advancedSyntax=true&analyticsTags=backend","processingTimeMS":11,"processingTimingsMS":{"_request":{"roundTrip":13},"fetch":{"query":7,"scanning":1,"total":9},"total":11},"query":"Chai Discovery","serverTimeMS":12}