frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Show HN: Buzz Bear – Monitor Reddit with Semantic Search ($0/mo cloud GPU costs)

https://www.buzzbear.ai/
1•yosalama•1m ago•0 comments

Show HN: Mantic.sh – Search 480k files in 0.46s without embeddings

1•marcoaapfortes•1m ago•1 comments

Moor 1.0-Beta1

https://timbran.org/announcing-moor-1-0-beta1.html
1•todsacerdoti•2m ago•0 comments

I Built Vissper Three Times Before Getting It Right

https://malvik.de/blog/vissper-oss
1•svenmalvik•3m ago•0 comments

Show HN: Query Wikidata with DuckDB Instead of Sparql

https://github.com/piebro/wikidata-extraction
1•piebro•3m ago•0 comments

Spectral Scanner – See who's tracking you in real-time

https://www.uniquenesslabs.com/browserscanner
1•Uniqueness_Labs•4m ago•1 comments

Show HN: I built a CLI-first pipeline that turns Wikipedia into narrated videos

https://github.com/NPgreatest/Wiki2Video
1•np123greatest•4m ago•1 comments

The Slow Death of Scaling (deep learning models)

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5877662
1•lairv•4m ago•0 comments

What managing 300+ Windows laptops really taught me

1•elinacorval•5m ago•0 comments

RFK Jr.'S Tuskegee Experiment

https://pauloffit.substack.com/p/rfk-jrs-tuskegee-experiment
1•doener•5m ago•0 comments

The Shift from Traditional SEO to AI Visibility

https://aiforcontentmarketing.ai/the-shift-from-traditional-seo-to-ai-visibility-navigating-the-2...
1•pakostina•6m ago•0 comments

Sleep patterns assessed by a smartphone app and work productivity loss

https://www.nature.com/articles/s41746-025-02155-3
1•PaulHoule•8m ago•0 comments

The decline and fall of the American technology industry

https://twitter.com/WillManidis/status/2008526902554775586
1•jger15•9m ago•0 comments

Show HN: Build a Directory in 7 Days with Vibe Coding (Free Roadmap)

https://directoryideas.ai/vibe-coded-roadmap
1•tejas3732•10m ago•0 comments

Making Games in Go with Ebitengine

https://alicegg.tech/2026/01/06/ebitengine
1•zer0tonin•10m ago•0 comments

Deep in the vaults: the Bank of England's £1.4B Venezuelan gold conundrum

https://www.theguardian.com/business/2026/jan/06/bank-of-england-venezuelan-gold-nicolas-maduro-u...
2•beardyw•12m ago•0 comments

Show HN: I'd never touched Swift. Built a Mac app in 4 weeks

https://www.notchie.app/
1•amortka•14m ago•0 comments

Microsoft didn't rebrand Office to Microsoft 365 Copilot

https://www.theverge.com/tech/856149/microsoft-365-office-rebrand-copilot
1•taubek•15m ago•0 comments

Ask HN: When tests keep passing but design stops moving

1•felixasher•17m ago•0 comments

Mad Skills to Learn the Universe

https://www.dotkam.com/2026/01/06/mad-skills-to-learn-the-universe/
1•tolitius•18m ago•1 comments

Deep Agents at Scale: The Three Problems You Can't Ignore

https://www.bolshchikov.com/p/deep-agents-at-scale
1•bolshchikov•18m ago•0 comments

Nvidia CEO reveals new 'reasoning' AI tech for self-driving cars

https://www.theguardian.com/technology/2026/jan/05/nvidia-chips-jensen-huang
1•pmg101•18m ago•0 comments

Statistical Rethinking 2026 by Richard McElreath [video]

https://www.youtube.com/watch?v=ztbYkBPDOgU&list=PLDcUM9US4XdNOlqSyhe38US8mFgmqzI14
1•smoe•19m ago•0 comments

Why German Strings Are Everywhere?

https://cedardb.com/blog/german_strings/
1•byt3h3ad•20m ago•0 comments

Show HN: Bootstrapper Roadmap – Building and exiting with an OSS stack

https://github.com/pierpaolo28/Awesome-Bootstrapper-Roadmap
1•zlatanmunutd10•21m ago•1 comments

Show HN: I built an AI that gives verifiable answers

https://getcitedai.com
1•collin1•22m ago•0 comments

Log2(3) and Log2(5)

https://www.johndcook.com/blog/2026/01/05/log2-of-3-and-5/
1•ibobev•23m ago•0 comments

In what way is MJML responsive?

1•bbyford•24m ago•0 comments

Show HN: Intellistant, a 10-50x faster C++ alternative 2 LangChain 4 AI agents

https://github.com/pooriayousefi/intellistant
1•pooriayousefi•24m ago•1 comments

There Is No New Aesthetics

https://borretti.me/article/there-is-no-new-aesthetics
2•ibobev•25m ago•0 comments
Open in hackernews

Xsight Labs E1 DPU a 64-Core Arm Neoverse N2 800G DPU

https://www.servethehome.com/this-is-the-xsight-labs-e1-dpu-a-64-core-arm-neoverse-n2-800g-dpu/
18•ksec•2d ago

Comments

ksec•2d ago
I know hardware doesn't usually get attention on HN but this DPU trend is fascinating to me and I finally understand what Jansen was saying about GPU and the next innovation is in networking with DPU.
Already__Taken•2d ago
I wish I knew hardware better but I'm coming to the idea that an xTB nvme and ~25gb/s PoE pcie card in my gaming PC would be a sic always on device. Lots of server roles really just need a tiny IoT device but storage/backup/media you're talking NAS and I just don't want a big NAS around all the time.

So an always on network storage in my PC that's sometimes-on. Between NVME-of and pcie that should be doable I would have though. PoE from my router that's got a tiny UPS and that's pretty sick.

Maybe Valve should make it so everyone with a gaming pc it's re-downloading their steam library or running 2 computers just to open a file.

wmf•2d ago
Basically you're talking about https://mikrotik.com/product/ccr2004_1g_2xs_pcie or https://www.solid-run.com/news/the-first-software-defined-dp... but I think a real NAS is cheaper.
Already__Taken•1d ago
Those are very cool thank you. Still need the host device to remain powered on however so not quite.
Veserv•2d ago
What exactly is that benchmark testing? Seems like a lot of cores to handle just 800 Gb/s; that is only ~12.5 Gb/s per core. Even with their ~19% headroom that is still only ~15 Gb/s per core. You can terminate encrypted transport multiple times faster than that and unencrypted transport 10x faster than that comfortably in software on regular cores.
wmf•2d ago
They might be doing 800+800 which would be 30 Gbps per core.

Ideally they would run some kind of cloud vSwitch benchmark that includes encryption (e.g. ESP), encapsulation (e.g. GENEVE), routing, stateful firewalling, NAT, monitoring (e.g. NetFlow/sFlow), etc. It's a lot of work.