A lightweight TypeScript library for assertion-based runtime data validation (github.com)

1 points by nayajunimesh 26s ago 1 comments

Omarchy Is Out (world.hey.com)

1 points by kristianp 7m ago 0 comments

Lithium May Be the Key to Stopping Alzheimer's (scitechdaily.com)

1 points by sowbug 9m ago 0 comments

Lightcap AI – Weighted Language Model Agent (lightcap.ai)

2 points by WASDAai 17m ago 1 comments

I hacked a way to use URL shortener to extract users' emails from GitHub/Docker

1 points by lexokoh 19m ago 0 comments

ThinkMesh: A Python lib for parallel thinking in LLMs (github.com)

2 points by martianlantern 20m ago 0 comments

Retailers slash prices of AMD and Intel latest EPYC and Xeon CPUs by up to 50% (tomshardware.com)

1 points by ksec 22m ago 1 comments

Mirage2.net AI World Generator

1 points by RyanMu 24m ago 0 comments

There's No Gold in Removing Signaling Costs (liorbd.me)

1 points by liorben-david 29m ago 0 comments

Kioxia's 5TB, 64 GB/s flash module puts NAND toward the memory bus for AI GPU (tomshardware.com)

1 points by ksec 32m ago 1 comments

Concept Poisoning: Probing LLMs without probes (lesswrong.com)

1 points by qouteall 33m ago 0 comments

I started building a permissions broker for AI agent for myself (github.com)

1 points by lexokoh 39m ago 0 comments

Show HN: brew-cleaner – CLI to bulk uninstall Homebrew formulae and free space (github.com)

2 points by theharshb 42m ago 0 comments

Google Develops KFuzzTest for Fuzzing Internal Linux Kernel Functions (phoronix.com)

2 points by westurner 45m ago 3 comments

LLMs and the Russellian Inversion (philosophicalhacker.com)

2 points by kmdupree 57m ago 0 comments

OmniNeural – First NPU-Aware Multimodal Model (huggingface.co)

1 points by huragok 57m ago 0 comments

Show HN: AWS Cost Explorer Clone (github.com)

1 points by 9woc 58m ago 0 comments

We Built This to Film What's 7,500 Light-Years Away [video] (youtube.com)

2 points by angilr 1h ago 0 comments

What Are Oklch Colors? (jakub.kr)

4 points by vinhnx 1h ago 1 comments

Show HN: Text-to-Explainer Video in Seconds (trytorial.com)

1 points by bames_jond 1h ago 1 comments

Show HN: How to Build a Coding Agent (free workshop) (ghuntley.com)

26 points by ghuntley 1h ago 3 comments

I guess I was wrong about AI persuasion (dynomight.substack.com)

2 points by paulpauper 1h ago 0 comments

Day 56 of latte art until I'm good it (instagram.com)

1 points by shreya51 1h ago 4 comments

You shouldn't salt a leech that's sucking your blood (cbc.ca)

1 points by pabs3 1h ago 0 comments

Kinds of Programming People (leftoversalad.com)

4 points by saulpw 1h ago 0 comments

I built a tiny mac app to monitor and manage my development processes (github.com)

18 points by lexokoh 1h ago 2 comments

PeaceFounder: Simple centralized Evoting thats voter-verifiable and private (peacefounder.org)

1 points by TheWiggles 1h ago 0 comments

Show HN: Run AI models directly in the browser – no server or internet required (private-ai-chat.vercel.app)

2 points by nadchif 1h ago 3 comments

Drawing Persona 5 UI with Compose [video] (youtube.com)

1 points by koakuma-chan 1h ago 0 comments

Scientists Are Caught in a Political Trap (theatlantic.com)

2 points by petethomas 1h ago 1 comments

Exceptional Nordic heatwave stumps tourists seeking shade (yahoo.com)

2 points by PaulHoule 1h ago 0 comments

CVE-2025-8901: OOB write in Chrome (nvd.nist.gov)

2 points by larkinrichards 1h ago 0 comments

Using information theory to solve Mastermind (goranssongaspar.com)

1 points by SchwKatze 1h ago 0 comments

Review: Ollantay (astralcodexten.com)

1 points by paulpauper 1h ago 0 comments

Agentic Browser Security: Indirect Prompt Injection in Perplexity Comet (brave.com)

6 points by drak0n1c 1h ago 2 comments

Lovable raises $200M and partners with Hangout for vibe coding with music (hangout.fm)

2 points by ljlolel 1h ago 1 comments

AI App Dev Log: The Story of Our App Begins

3 points by monstergpt 1h ago 0 comments

Claude-Trace (github.com)

3 points by handfuloflight 2h ago 0 comments

What Is Customer Identity and Access Management (CIAM)? CIAM Explained (fusionauth.io)

1 points by mooreds 2h ago 0 comments

How A Neuron Learns (rvns.moe)

2 points by jxmorris12 2h ago 0 comments

Scientists found the missing nutrients bees need – Colonies grew 15-fold (sciencedaily.com)

2 points by fcpguru 2h ago 2 comments

Deal to get ChatGPT Plus for whole of UK discussed by Open AI boss and minister (theguardian.com)

3 points by thunderbong 2h ago 0 comments

Lasersaur – open-source laser cutter (lasersaur.cc)

1 points by QuantumNomad_ 2h ago 0 comments

See a Satellite Tonight (james.darpinian.com)

2 points by simonpure 2h ago 0 comments

Low Earth Orbit Visualization (platform.leolabs.space)

2 points by simonpure 2h ago 0 comments

Joe Gebbia Appointed Chief Design Officer of USA (twitter.com)

1 points by drak0n1c 2h ago 0 comments

Neuromuscular Aim Assist (youtube.com)

2 points by sakopov 2h ago 0 comments

Lightcap – Scientific AI Chatbot (wasda.ai)

1 points by WASDAai 2h ago 2 comments

Great Principles of Computing (denninginstitute.com)

3 points by pillars 2h ago 0 comments

Improving Our Nation Through Better Design (whitehouse.gov)

4 points by efavdb 2h ago 0 comments

AMD Ryzen AI Max+ 395 vs. Ryzen 9 9950X vs. Ryzen 9 9950X3D Linux Performance

15 teekert 8 8/24/2025, 12:02:32 AM phoronix.com ↗

Comments (8)

jauntywundrkind · 3h ago

Feels likely to be a pretty direct test of memory bandwidth. Strix Halo (the AI Max+ 395) has quad ddr5-8000, which is both twice the width and more mhz. This is a great opportunity to see what workloads will make use of the 3d-vcache too!

I'm personally pretty excited for a new round of Mobile on Desktop. Usually a pretty good % of desktop speed, but tuned for much lower power, and the price is usually a little better. I have the Minisforum 795s7 with the 7945hx, which is a delightful zen4 that's very affordable. https://www.tomshardware.com/pc-components/cpus/amd-unleashe...

wmf · 2h ago

I don't know if the CCDs can actually use that much bandwidth. It's really for the GPU.

jauntywundrkind · 49m ago

There were sometimes quite notable differences between 32-core ThreadRipper at quad channel ddr5 6400mhz vs 4800mhz. https://www.phoronix.com/review/threadripper-9000-ddr5-6400-...

Indeed yes it seems like the dual channel 6000MHz wasn't the limiting factor. But also the lower power 395 has a couple massive wins, and I suspect it was almost entirely the memory bandwidth. For example the LLM test on page 2: it was that memory bandwidth making that colossal difference.

In a lot of tests, it seems like the extra 100w the 9950's were getting probably helped a lot!

wmf · 2h ago

Terrible benchmarks. They should probably compare the 395 GPU against the 9950 CPU or against a 7600/9060. If you're just going to use the CPU obviously the 395 is not what you want.

kcb · 2h ago

I don't think that was the point of these benchmarks. Instead it is interesting to see how memory bandwidth effects CPU performance.

wmf · 2h ago

These benchmarks are measuring TDP not memory bandwidth.

yogorenapan · 2h ago

I hate the naming so much. I cannot tell at all which is meant to be newer or superior. It's also an odd benchmark considering the LLM should be running on the GPU with shared memory rather than CPU

wtallis · 1h ago

Intel and AMD have pretty much never made it straightforward to compare laptop part numbers against desktop part numbers and know what's faster or newer. That's not a new problem.

If you compare the Ryzen AI Max+ 395 against the other AMD mobile parts in their current lineup, it's pretty clear where it ranks. And for however long they're using a new and different naming scheme for their mobile parts while still using the old naming scheme for their desktop parts, there's the distinct and rare advantage that the model numbers are not misleading about how a high-end mobile part ranks against desktop parts.