frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

US July gaming hardware spending dropped to near-2020 out-of-stock lows

https://www.eurogamer.net/circana-july-2026-hardware-software-sales
2•haunter•3m ago•0 comments

The Threat of AI to Media Archives

https://ebird.org/news/threat-of-ai-to-archives
2•steptwo•4m ago•0 comments

You Will Go To Mars (a picture story book)

https://www.youwillgotomars.com/
1•rtills•4m ago•0 comments

Robots are asking for tips. But who gets the money?

https://www.bbc.com/future/article/20260818-do-you-have-to-tip-the-robots
2•rmason•5m ago•1 comments

How to Destroy AI Scam Phone Calls (Short Video)

https://www.youtube.com/shorts/aa0v-24EFy0
1•rmason•8m ago•0 comments

Early-life stress leaves a 'scar' inside brain cells

https://medicine.washu.edu/news/how-early-life-stress-leaves-a-scar-inside-brain-cells/
2•gmays•8m ago•0 comments

Single-User Inference Card

https://markohaberl.substack.com/p/single-user-inference-card
1•mhaberl•10m ago•0 comments

The Jason Arday delusion: Scapegoating is punitive and crass

https://unherd.com/2026/08/the-jason-arday-delusion/
2•cahlingl•11m ago•0 comments

AI Is Revolutionizing Strategic Decision-Making

https://hbr.org/2026/09/ai-is-revolutionizing-strategic-decision-making
1•devy•11m ago•0 comments

Context-Aware Inlining: Using Call-Stack Profiles for Fast and Smaller Binaries

https://dl.acm.org/doi/10.1145/3830016
1•matt_d•12m ago•0 comments

Everybody watches some TV

https://profane-tmesis.info/everybody-watches-some-tv
1•riskone•13m ago•0 comments

The Wonders of the Male Human Pelvis

https://nautil.us/the-wonders-of-the-male-human-pelvis-1283947
2•littlexsparkee•16m ago•0 comments

FTC says "personalized pricing" based on consumer data could violate the law

https://www.cbsnews.com/news/ftc-personalized-pricing-warning-consumers/
5•rawgabbit•16m ago•0 comments

ICE and Iran worked together to deport Iranians from U.S.

https://www.pbs.org/newshour/world/ice-and-iran-worked-together-to-deport-iranians-from-u-s-newly...
3•cocacola1•16m ago•0 comments

We published 421 AI-written articles in five months. They earned 10 clicks

https://blog.marketingsohigh.com/421-ai-articles-10-clicks-what-the-data-showed/
1•csroay•16m ago•0 comments

So Many Fisker Ocean in New York, but Why?

https://thelastdriverlicenseholder.com/2025/05/13/so-many-fisker-ocean-in-new-york-but-why/
3•eric_h•17m ago•1 comments

Show HN: Molecule.dev is an AI-first composable package ecosystem

https://github.com/molecule-dev/molecule
1•loh•17m ago•1 comments

Protocol-Aware Deterministic Simulation Testing

https://tigerbeetle.com/blog/2026-08-20-protocol-aware-dst/
1•sebg•17m ago•0 comments

Bradley Cadenhead: The Boy Who Became 764

https://joniejohnstonpsyd.substack.com/p/bradley-cadenhead-the-boy-who-became
1•Teever•19m ago•0 comments

Mark Buys a Castle

https://techcrunch.com/2026/08/20/mark-buys-a-castle/
3•babelfish•20m ago•1 comments

Fantasy Map Generator

https://azgaar.github.io/Fantasy-Map-Generator/
1•mooreds•20m ago•1 comments

Show HN: Praxos – team messaging with built-in memory

https://www.praxos.ai/
2•mogusian•21m ago•1 comments

The Generative AI Learning Penalty: Evidence from Chinese Secondary Education

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618
3•EA-3167•21m ago•0 comments

Chinese Video Platform Bilibili Is Going Global

https://www.engadget.com/2241074/chinese-video-platform-bilibili-is-going-global/
2•rdmuser•24m ago•0 comments

Opus: A Minimal Statically-Scoped Lisp Dialect Based on F-Expressions

https://github.com/yaoshiu/opus
1•birdculture•24m ago•0 comments

Ameliorate

https://ameliorate.app/
2•hakkikonu•24m ago•0 comments

Teaching a local LLM to reason about a new domain through continued pretraining

https://www.teachmecoolstuff.com/viewarticle/teaching-a-local-llm-a-new-domain
2•dev-experiments•25m ago•0 comments

Show HN: PingChange – monitor web pages and OpenAPI specs for meaningful changes

https://pingchange.com
1•epigos•25m ago•0 comments

RoboDojo Scores 32 of 34 Robot Policies Under 10% Success

https://topicqueue.substack.com/p/robodojo-scores-32-of-34-robot-policies
1•DISCURSIVE•25m ago•0 comments

A breakdown of the August 2026 GTA 6 gameplay leaks

https://github.com/zyrexdz/cyberleek-leak-research
1•oliverkwebb•27m ago•0 comments
Open in hackernews

Show HN: Praxos – team messaging with built-in memory

https://www.praxos.ai/
2•mogusian•21m ago
Hey HN! What does a regular workflow at a startup look like?

Probably like this:

> One founder talks to customers

> Another founder builds

> A customer asks for something. The first founder explains it to the second. The second founder explains it to Claude

> Two weeks any one of the founders asks one of the following questions: * Why are we building it this way? * What happened here? * What did we promise ACME we’d do? * How far along are we on X?

> Then you scream on the inside because you have to figure out what happened...

Before agents, the person who did one part of the work usually held a mental map. They could share the context with other people. Increasingly, they can’t.

Small teams can run far more work in parallel than ever before thanks to AI, but someone now has to connect all that context manually. The reasoning behind one product change might now be scattered across a customer call, two emails, a Slack thread, and four Claude conversations.

Praxos is a team messaging app.

It has channels, groups and threads like you’d expect. However, places where work happens—Github, harnesses, email, calls—get linked and converted into a living record of your company. This means everyone is always up to date with the latest developments.

Agents can join conversations and inherit the same context humans do.

Here are examples we’ve seen in production:

You talk to a customer. Later, when an engineer or coding agent works on the feature, your conversation and the customer’s information are already in context. A new engineer picks up an existing project. They have immediate access to info about who worked on what part of the code base, the deltas across changes, and a granular view of their conversations with coding agents. There’s no need to interrogate the CTO. Three weeks later, someone asks why the engineer made a certain decision about the feature. Praxos fetches the engineer’s convos, the original feature request and call logs between the founder and the engineer. Then it contextualizes the work.

We’re early, but the core product already works and is in use. You get messaging, search, memory and connectors. You can bring your own agents, use ours, or both.

A desktop app is going live next week. We’re building native calls and mobile apps, so that all the communications that go into a company’s history can happen on Praxos.

If you’re a 2 to 10 person team where one person is spending all day with customers while the rest of the team is shipping with Claude/Codex, we’d love to personally onboard you.

Especially if you’ve already caught yourself asking “what did the customer want?” or “what’s going on with X?” last week.

The first 10 teams to sign up get Praxos for free for 6 months.

Email lucas@praxos.ai to get started!

Comments

tidbeck•11m ago
May I ask how you solved iMessage and WhatsApp as they do not allow bots as I understand it.