frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Micro-Front Ends in 2026: Architecture Win or Enterprise Tax?

https://iocombats.com/blogs/micro-frontends-in-2026
1•ghazikhan205•2m ago•0 comments

Japanese rice is the most expensive in the world

https://www.cnn.com/2026/02/07/travel/this-is-the-worlds-most-expensive-rice-but-what-does-it-tas...
1•mooreds•2m ago•0 comments

These White-Collar Workers Actually Made the Switch to a Trade

https://www.wsj.com/lifestyle/careers/white-collar-mid-career-trades-caca4b5f
1•impish9208•2m ago•1 comments

The Wonder Drug That's Plaguing Sports

https://www.nytimes.com/2026/02/02/us/ostarine-olympics-doping.html
1•mooreds•2m ago•0 comments

Show HN: Which chef knife steels are good? Data from 540 Reddit tread

https://new.knife.day/blog/reddit-steel-sentiment-analysis
1•p-s-v•3m ago•0 comments

Federated Credential Management (FedCM)

https://ciamweekly.substack.com/p/federated-credential-management-fedcm
1•mooreds•3m ago•0 comments

Token-to-Credit Conversion: Avoiding Floating-Point Errors in AI Billing Systems

https://app.writtte.com/read/kZ8Kj6R
1•lasgawe•3m ago•1 comments

The Story of Heroku (2022)

https://leerob.com/heroku
1•tosh•3m ago•0 comments

Obey the Testing Goat

https://www.obeythetestinggoat.com/
1•mkl95•4m ago•0 comments

Claude Opus 4.6 extends LLM pareto frontier

https://michaelshi.me/pareto/
1•mikeshi42•5m ago•0 comments

Brute Force Colors (2022)

https://arnaud-carre.github.io/2022-12-30-amiga-ham/
1•erickhill•8m ago•0 comments

Google Translate apparently vulnerable to prompt injection

https://www.lesswrong.com/posts/tAh2keDNEEHMXvLvz/prompt-injection-in-google-translate-reveals-ba...
1•julkali•8m ago•0 comments

(Bsky thread) "This turns the maintainer into an unwitting vibe coder"

https://bsky.app/profile/fullmoon.id/post/3meadfaulhk2s
1•todsacerdoti•9m ago•0 comments

Software development is undergoing a Renaissance in front of our eyes

https://twitter.com/gdb/status/2019566641491963946
1•tosh•9m ago•0 comments

Can you beat ensloppification? I made a quiz for Wikipedia's Signs of AI Writing

https://tryward.app/aiquiz
1•bennydog224•10m ago•1 comments

Spec-Driven Design with Kiro: Lessons from Seddle

https://medium.com/@dustin_44710/spec-driven-design-with-kiro-lessons-from-seddle-9320ef18a61f
1•nslog•10m ago•0 comments

Agents need good developer experience too

https://modal.com/blog/agents-devex
1•birdculture•12m ago•0 comments

The Dark Factory

https://twitter.com/i/status/2020161285376082326
1•Ozzie_osman•12m ago•0 comments

Free data transfer out to internet when moving out of AWS (2024)

https://aws.amazon.com/blogs/aws/free-data-transfer-out-to-internet-when-moving-out-of-aws/
1•tosh•13m ago•0 comments

Interop 2025: A Year of Convergence

https://webkit.org/blog/17808/interop-2025-review/
1•alwillis•14m ago•0 comments

Prejudice Against Leprosy

https://text.npr.org/g-s1-108321
1•hi41•15m ago•0 comments

Slint: Cross Platform UI Library

https://slint.dev/
1•Palmik•19m ago•0 comments

AI and Education: Generative AI and the Future of Critical Thinking

https://www.youtube.com/watch?v=k7PvscqGD24
1•nyc111•19m ago•0 comments

Maple Mono: Smooth your coding flow

https://font.subf.dev/en/
1•signa11•20m ago•0 comments

Moltbook isn't real but it can still hurt you

https://12gramsofcarbon.com/p/tech-things-moltbook-isnt-real-but
1•theahura•24m ago•0 comments

Take Back the Em Dash–and Your Voice

https://spin.atomicobject.com/take-back-em-dash/
1•ingve•24m ago•0 comments

Show HN: 289x speedup over MLP using Spectral Graphs

https://zenodo.org/login/?next=%2Fme%2Fuploads%3Fq%3D%26f%3Dshared_with_me%25253Afalse%26l%3Dlist...
1•andrespi•25m ago•0 comments

Teaching Mathematics

https://www.karlin.mff.cuni.cz/~spurny/doc/articles/arnold.htm
2•samuel246•28m ago•0 comments

3D Printed Microfluidic Multiplexing [video]

https://www.youtube.com/watch?v=VZ2ZcOzLnGg
2•downboots•28m ago•0 comments

Abstractions Are in the Eye of the Beholder

https://software.rajivprab.com/2019/08/29/abstractions-are-in-the-eye-of-the-beholder/
2•whack•28m ago•0 comments
Open in hackernews

Show HN: I built an AI SuperConnector for founders and VCs

https://www.vance.so/
4•yednap868•3mo ago
Hey everyone,

I built Vance, an AI SuperConnector that makes warm introductions between founders and investors (and eventually anyone who needs the right connection).

You just tell Vance what you’re looking for - e.g. “I’m raising for my seed round” or “Looking for a co-founder in AI” and he finds the right person from your extended network to introduce you to.

Right now, we’re experimenting with curated networks like IITs, IIMs, and founder communities to help Vance build context and trust.

I’d love feedback from this community -

1. Is this something you’d use? 2. How do you think about privacy/trust in AI led intros? 3. Any thoughts on building strong network density early on?

Would really appreciate any suggestions or brutally honest feedback

Here’s the link if you want to try it: https://vance.so

Comments

rluther5•3mo ago
I get the pitch, but this feels like yet another attempt to automate something that fundamentally only works because it’s human.

Warm intros aren't just about “matching nodes on a graph.” They’re about credibility and skin in the game. When an actual person introduces you, they’re implicitly staking their reputation. An AI “superconnector” can’t do that — at least not in a way investors or founders will trust when the stakes are real.

Also, the trend of turning personal networks into monetizable data graphs is getting pretty dystopian. We keep saying “privacy matters” while happily piping personal+contextual trust data into the next startup. I’m sure the intentions are good, but this is the kind of tool that looks harmless until it isn’t.

HN crowd will hate me for this, but if your startup only works when you get early density from elite gated networks (IIT/IIM/YC/etc.), the “ democratizing access to opportunity” narrative rings a little hollow.

Not saying this can’t become useful — but I’d want to see:

Crystal-clear privacy commitments Evidence that investors actually trust AI-mediated intros Proof it works beyond curated privilege clusters

Right now it feels closer to automating the illusion of warm intros than the real thing.

yednap868•3mo ago
That’s a really fair take and honestly, I agree with a lot of what you said.

Warm intros work because they’re human. When someone introduces you, they’re putting their reputation behind it. Vance isn’t trying to replace that, it’s more like helping people make those intros faster and smarter. The AI just understands context and timing, the human still approves the intro.

On privacy, completely get the concern. We don’t scrape or sell data. Everything is opt-in through alumni or founder networks that choose to partner. It’s less about monetizing networks and more about making them actually useful.

And yeah, starting with IIT/IIM-type clusters is just to prove that the model works where trust density is high. The long-term goal is to open this up so access to good intros isn’t a privilege.

Appreciate you sharing this, these are exactly the kind of questions that help us build this right.

garkotipankaj•3mo ago
On first glance, this seems very gimmicky. The whole wolf fursona thing does not make sense. What's the logic here? That wolves are social animals? I digress.

There is nothing on the website that works to remove this initial skepticism. Your AI maybe actually wonderful but atleast for me, it is hard to trust it - it is purely because you have chosen to present it in a way that is basically a cosplay. If I trust an AI to connect me with VCs and potential employers / clients, I definitely expect it to be a lot more serious in presentation. Right now it looks like a furry's personal website. No shade to furries but it simply does not win my trust. Please re-consider the way you are branding this if you want people to use it like a serious tool.

Having said that, the core idea seems interesting to me. LinkedIn has become a cesspool of AI slop. People writing posts with AI and then people commenting on those posts with AI. Funny enough, AI itself seems to be the thing that will end up saving us from that non-sense.

I was going to try this but I do not have WhatsApp. Most people in North America don't. So I can't even access your product. Thanks for wasting my time.

yednap868•3mo ago
That’s actually great feedback. The “wolf” angle was meant to make Vance feel a bit more personal and less like another cold SaaS tool, but I get how it can come off as gimmicky or even confusing if you’re seeing it for the first time. Appreciate you calling that out.

The idea is to build something that actually feels alive, not a form or dashboard but yeah, totally hear you on the trust side. We’ll definitely tone down the fluff and focus on clarity. And you’re right about WhatsApp, it’s mainly because we started in India where it’s the default messaging app. We’re thinking of giving users the ability to call Vance directly and reach out on LinkedIn.

Appreciate you checking it out (even if it wasted a few minutes). Thanks for being honest.

vidyanshu•3mo ago
1. Cool but very ambitious idea. Will I use it? Yes, if it really works. But I'm not really optimistic about Vance working because when humans intro someone, they’re putting reputation on the line which in this case Vance don't have. 2. The explanation on the applications about not storing or selling data helps, but in Vance's case, it might be worth surfacing how context is inferred, e.g is the matching based on declared intent ("I’m raising for X") or behavioural data (if yes, how are you getting it?)
yednap868•3mo ago
Fair point and you’re right, human intros carry reputation, that’s the real trust layer we’re trying to rebuild in an AI first way.

Vance starts small - making intros that actually lead somewhere, and over time, builds his own kind of reputation based on outcomes. That’s the part I’m most excited about. Also, totally agree on data, Vance doesn’t store or sell any. Context purely comes from what users share - intent, story, timing not behavioural tracking.

Curious though - how would you imagine trust being built for something like this?