frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Power tools got worse on purpose

https://www.worseonpurpose.com/p/your-power-tools-got-worse-on-purpose
1•longhaul•54s ago•0 comments

A New Chapter for Ruby Central

https://rubycentral.org/news/a-new-chapter-for-ruby-central/
1•campuscodi•2m ago•0 comments

Quantum Computers Are Not a Threat to 128-Bit Symmetric Keys

https://words.filippo.io/128-bits/
1•hasheddan•2m ago•0 comments

Show HN: Open-source alternative HN front page with point highlights and search

https://github.com/pretzelai/hackernewsx
1•ramonga•5m ago•0 comments

Show HN: Ctx – a /resume that works across Claude Code and Codex

https://github.com/dchu917/ctx
2•dchu17•5m ago•0 comments

The Bacterial Flagellar Motor: What Physical Life Force Turns Biology's Wheels?

https://www.quantamagazine.org/what-physical-life-force-turns-biologys-wheels-20260420/
1•ganitam•5m ago•0 comments

We Accepted Surveillance as Default

https://vivianvoss.net/blog/why-we-accepted-surveillance
5•speckx•5m ago•0 comments

China to break U.S. reliance after uncovering €1.4B cache of ultra-pure quartz

https://www.leravi.org/china-poised-to-break-final-u-s-reliance-after-uncovering-e1-4-billion-cac...
2•frasermarlow•6m ago•0 comments

Why you should work for a top tier tech company

https://jsavage.xyz/2026/04/20/why-you-should-work-for-a-top-tier-tech-company/
2•JSavageOne•7m ago•0 comments

Humanoid robots show rapid advances racing past humans in Beijing half-marathon

https://www.theguardian.com/sport/2026/apr/19/humanoid-robots-race-beijing-half-marathon
1•KolmogorovComp•8m ago•0 comments

Let It Slop: A New Approach to Modularity in the Age of AI Code Generation

https://tomash.wrug.eu/blog/2026/04/19/new-modularity/
1•lackoftactics•8m ago•0 comments

Distributed Terminal Game for 24 Players (No Server)

https://github.com/ObliviousCompute/ObliviousCompute/blob/main/Byzantium/README.md
1•InfiniteMass•8m ago•1 comments

Ormah – Collective Memory Layer that whispers to your agents [video]

https://www.youtube.com/watch?v=IngB55jdnlc
1•rajit2•8m ago•1 comments

Random thoughts while gazing at the misty AI Frontier

https://blog.eladgil.com/p/random-thoughts-while-gazing-at-the
1•gmays•9m ago•0 comments

Show HN: Make an immutable web flipbook from any doc, video or browser replay

https://browser-session-8faya.pages.dev
1•keepamovin•9m ago•0 comments

Ukraine's Second Miracle Year

https://www.thebulwark.com/p/ukraines-second-miracle-year-putin-trump-zelensky-war-drones-oil-exp...
2•kaycebasques•9m ago•0 comments

General Health Checks Are Harmful

https://brownstone.org/articles/general-health-checks-are-harmful/
1•nradov•10m ago•0 comments

Epistemic Suicide in AI: How Binary Feedback Distorts Model Reasoning

https://medium.com/@erinacius4455/epistemic-suicide-in-ai-how-binary-feedback-quietly-destroys-re...
1•alex_gold•10m ago•0 comments

Artemis II: Why Going Back to the Moon Is a Big Deal

https://www.akashtandon.in/interactive-explainers/artemis-ii/
1•akashtndn•11m ago•0 comments

Context Security Response Statement

https://context.ai/security-update
2•firloop•13m ago•0 comments

EV sales soar in main European markets as drivers shun expensive petrol

https://www.reuters.com/sustainability/climate-energy/ev-sales-soar-main-european-markets-drivers...
3•akyuu•14m ago•1 comments

Show HN: Making video games every day with Claude (Day 5: Minefield)

https://gamevibe.us/5-minefield
1•pzxc•14m ago•0 comments

Software Engineering Practices (Are Also) Useful for Token Reduction

https://robotpaper.ai/software-engineering-practices-are-also-useful-for-token-reduction/
2•speckx•15m ago•0 comments

Hustlers are cashing in on China's OpenClaw AI craze

https://www.technologyreview.com/2026/03/11/1134179/china-openclaw-gold-rush/
1•omer_k•16m ago•0 comments

Release of Writing App: Alistair v1.0

https://alistaircrowe.com/
2•RyanProietto•17m ago•0 comments

Meta's Smart Glasses Dominate Wearables Market in 2026

https://www.techbuzz.ai/articles/meta-s-smart-glasses-dominate-wearables-market-in-2026
1•Vaslo•18m ago•0 comments

Not Your Father's Internet

https://systemsapproach.org/2026/04/20/not-your-fathers-internet/
1•zdw•18m ago•0 comments

Online response to the attack on Sam Altman's house shows a generational divide

https://fortune.com/2026/04/14/ai-backlash-revolutionary-sam-altman-molotov-cocktails-data-centers/
1•measurablefunc•18m ago•0 comments

Postmortem-Driven Development

https://infisical.com/blog/postmortem-driven-development
3•vmatsiiako•19m ago•0 comments

Why every developer needs their own agent-skills (and how I code in 04/2026)

https://olshansky.substack.com/p/why-every-developer-needs-their-own
1•Olshansky•19m ago•0 comments
Open in hackernews

Show HN: Made a highly organised email client with prompt-free AI within

1•neerajnathany•1h ago
Hey HN, I've been working on a problem that had me annoyed since a long while. The fact that our email continues to be unorganized, cluttered, and fatiguing even today, is quite surprising tbh. Even the "modern" clients today simply put our emails in a table, support a few keyboard shortcuts, a few simple features, and are done. The newer ones provide AI-based compose assist, and the newest ones now want us to prompt for everything! That seems backwards. By the time you'll type "find that marketing agency's latest invoice from last Tuesday," you would've already spent more time and effort than you should've needed to.

There is still *no higher order classification, no deeper grouping*. Every email — a project update, a transaction, a human message; appears the same. Key attachments are always elusive when needed. *Trivial auto-updates crowd us*, and *long forwarded threads are still painful to follow with the untidy nature of their chaining.*

I'm sure I am not the only one who found this irritating!

So I built Faraday to finally solve email for good. It uses a classification and extraction pipeline on every incoming email before it has even arrived for the user. *No prompts, no triggers.* It does 3 things automatically:

*1. Higher-order classification:* Is this a Linkedin update, a transaction, a human conversation, an OTP, a booking confirmation? Not just "primary vs promotions" — actual semantic classification over ~30 categories, sub-groups, statuses and even genres. And then the sub-categories are ordered basis what's best for them. Who cares about the chronology within 40 newsletters from last week. The genres and the brands across them are more important.

*2. Contextual extraction:* Relevant content in most emails is roughly just 12% of the total text. It extracts that (amounts, dates, names, actions, codes - different for different email types) and surface it at the top, so you get the meat upfront.

*3. Thread reconstruction: *Email threads are just terribly nested blockquotes. Faraday reconstructs and reassembles them into a clean, ordered conversation tree. This one is really quite slick.

A bunch of things were difficult to do here. Making all three layers work simultaneously across the spread of our email content, at inbox speeds, with optimum resource utilization, without a single training signal from the user. No setup or onboarding. To work as soon as you login. Super tricky to build.

Meanwhile, also ensured that *it is privacy-first*, doubling down on the best possible security and encryption standards (AES 256) right from the onset.

There are genuinely some nifty technical innovations in Faraday (enough to have even filed 2 patents for them:)

It works on top of all Gmail and Outlook accounts. It's live, compliance-approved, out-of-beta and a lot of people (who were earlier hooked on superhuman, spark) are now using this :)

Happy to go deep on any part of the architecture that you find interesting. Do try it: https://faraday.email (2-week free trial, and much inexpensive after)

Comments

slava_•1h ago
do I understand it correctly: you (1) have 30 pre-defined classes of emails, (2) you predict a class, (3) put email in a specific UI? I think dynamic clusters would be more helpful! every inbox is different. what if you would cluster my inbox and predict a cluster for a new email? you can do that based on the content and interactions.
neerajnathany•47m ago
The groups/topics are indeed dynamic! A slightly different set for each user. I wrote ~30 because those many are the categories + subcategories a user's emails roughly spread across. And it's still simple and intuitive to browse over:)