frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Personal AI Integrated with WhatsApp and Telegram

4•mogusian•2h ago
Hi all,

I wanted to share something I’ve been working on which has been super helpful to me already.

I’m the founder of a tech startup called Praxos, and over the past months I’ve started to struggle juggling all the different tasks and commitments falling on my plate. Often times I’ve caught myself spending most of my day chasing after personal life or mundane job work, only to realize that I couldn’t really (1) move the needle on things that mattered for my company’s success on that day; and that (2) I had a few hours left on my day but was too drained to do anything meaningful.

I am very organized and have a great memory, so I’ve never really used checklists or kept many things on the calendar. I understand that these things would’ve helped me prioritize better, but always felt that the hassle of putting things in place or keeping them updated outweighed the ease or just keeping them in my own memory. That said, I’ve had more and more “oh sh!t” moments lately with something important falling through the cracks.

Up until a month ago, the startup I co-founded with my best friend was in insurance. Our main bread and butter was being able to extract insurance data from anything and put it into anything. But as our systems got better, we started experimenting with using it for other things. For me, I realized that connecting the same email/calendar automations we built for insurance brokers, as well as the same capabilities to take my emails, documents, etc. and extract the data it needs to actually do stuff for me, can already make something that helps me.

Think about a simple WhatsApp or Telegram message that reminds me about lunch with a client, or that I have to run an errand, or having it find the blood pressure numbers I have somewhere my emails that the doctor keeps asking for to give me the refill and sending it to her already just made me feel less like I was balancing on a tight rope.

This was kind of a big moment for me, because we realized that having something which for us, ourselves, would be usable is so much more exciting that building a product in insurance. So we spent the past month taking what we had, and building it into exactly what we wanted to see.

For me, just the fact of knowing I can send it a random voice note on WhatsApp before bed, and having it immediately start coordinating with a client over email a time to meet next week has been a godsend.

So today, I’m sharing what I built here, hoping that people benefit from it.

We’re still actively fleshing it out, and we’re very open to hearing ideas, or things that people need. The goal’s to build a version of Siri / Android Assistant that can both (1) remember everything about you, and (2) do work for you. All of this one text / voice msg away.

If you’d like to check it out, here’s our site: https://mypraxos.com.

Speculative cascades – A hybrid approach for smarter, faster LLM inference

https://research.google/blog/speculative-cascades-a-hybrid-approach-for-smarter-faster-llm-infere...
1•emschwartz•2m ago•0 comments

Replit Agent 3

https://replit.com/agent3
1•tkgally•4m ago•0 comments

Chess.com co-founder discusses growing up in a religious cult

https://www.theguardian.com/lifeandstyle/2025/sep/07/chess-prodigy-trapped-religious-cult-danny-r...
1•gaws•5m ago•0 comments

Why our website looks like an operating system

https://posthog.com/blog/why-os
1•bnc319•8m ago•0 comments

The Big LLM Architecture Comparison by Sebastian Raschka [video]

https://www.youtube.com/watch?v=rNlULI-zGcw
1•tanelpoder•15m ago•0 comments

Albania appoints AI bot as minister to tackle corruption

https://www.reuters.com/technology/albania-appoints-ai-bot-minister-tackle-corruption-2025-09-11/
5•itchyjunk•16m ago•0 comments

Americans' knowledge of civics increases: survey

https://www.annenbergpublicpolicycenter.org/americans-knowledge-of-civics-increases-annenberg-sur...
2•hhs•19m ago•0 comments

State of Vibe Coding [pdf]

https://v0.app/vibecoding
1•twapi•21m ago•0 comments

US Initial Jobless Claims Jump to Highest in Almost Four Years

https://www.bloomberg.com/news/articles/2025-09-11/us-initial-jobless-claims-jump-to-highest-in-a...
3•JumpCrisscross•23m ago•0 comments

Four Steps to the Apocalypse: A talk at 60th anniversary of MIT's Project Mac [video]

https://www.youtube.com/watch?v=oDwlGSjICSQ
1•pcfwik•23m ago•0 comments

NASA Mars rover Perseverance finds best potential evidence yet of ancient life

https://www.abc.net.au/news/2025-09-11/nasa-discovery-ancient-life-perseverance/105760456
1•coolcoder613•23m ago•0 comments

AI Will Not Make You Rich

https://joincolossus.com/article/ai-will-not-make-you-rich/
1•fallmonkey•29m ago•0 comments

OpenAI Restructuring to Give Nonprofit $100B Stake

https://www.bloomberg.com/news/articles/2025-09-11/microsoft-openai-reach-early-agreement-in-for-...
2•mfiguiere•37m ago•0 comments

Proposal: SIMD: architecture-specific SIMD intrinsics

https://github.com/golang/go/issues/73787
2•kristianp•37m ago•0 comments

Mexican Cartels launder money with cryptocurrencies, concerts and timeshares

https://english.elpais.com/international/2025-08-30/mexican-cartels-new-money-laundering-business...
5•PaulHoule•37m ago•0 comments

Senator demands to know status of 'duplicate' SSA database 'immediately'

https://www.theregister.com/2025/09/11/ssa_doge_whistleblower_demand/
17•raw_anon_1111•39m ago•3 comments

Microsoft CEO Nadella admits company has to rebuild trust with employees

https://www.cnbc.com/2025/09/11/microsoft-ceo-nadella-says-company-must-rebuild-trust-with-employ...
6•cebert•40m ago•0 comments

'Microbial piracy' uncovers new way to fight drug-resistant infections

https://www.imperial.ac.uk/news/268213/microbial-piracy-uncovers-fight-drug-resistant-infections/
3•hhs•40m ago•0 comments

OpenAI and Microsoft MOU

https://twitter.com/OpenAINewsroom/status/1966253248886747576
2•itchyjunk•41m ago•0 comments

Automating multi-language SDK doc generation with testable code snippets

https://docs.hatchet.run/blog/automated-documentation
1•abelanger•41m ago•0 comments

Evaporating Water with Light

http://lll.scrybble.ink/Evaporating-water-with-light
1•azeirah•42m ago•0 comments

The Learning Zone Matrix

https://pointc.co/thelearningzone/
2•benwerd•43m ago•0 comments

Netanyahu is an unhinged narcissist who has sabotaged world peace

https://www.independent.co.uk/news/world/middle-east/netanyahu-qatar-israel-bombing-airstrike-doh...
13•NomDePlum•45m ago•0 comments

Exposure of Sensitive Information to an Unauthorized Actor in Argo-CD

https://radar.offseq.com/threat/cve-2025-55190-cwe-200-exposure-of-sensitive-infor-c3aae4b1
2•gpi•46m ago•0 comments

PsiQuantum raise 1B to build quantum computer in collab with Nvidia

https://quantumcomputingreport.com/psiquantum-raises-1-billion-to-build-fault-tolerant-quantum-co...
1•gsf_emergency_2•50m ago•0 comments

NVIDIA Accelerated IO (XLIO)

https://docs.nvidia.com/networking/display/xliov3312/xlio+library+architecture
2•tanelpoder•51m ago•0 comments

The Power of Ten: Rules for Developing Safety Critical Code [pdf]

http://spinroot.com/gerard/pdf/P10.pdf
2•ntnbr•51m ago•0 comments

Improving state machine code generation

https://trifectatech.org/blog/improving-state-machine-code-generation/
1•jrpelkonen•51m ago•0 comments

The bloat of edge-case first libraries

https://43081j.com/2025/09/bloat-of-edge-case-libraries
1•thunderbong•52m ago•1 comments

Evaluating and Optimizing LLM Applications with DSPy

https://pedramnavid.com/blog/dspy-evals/
1•pinkbeanz•53m ago•0 comments