Sign in

Scott Wasson

@scottwasson.bsky.social
638 followers 460 following 3.2K posts

Founder, The Tech Report and Ars Technica. Formerly AMD Radeon and Intel GPU groups and MLPerf at MLCommons.

PostsRepliesMedia
Scott Wasson @scottwasson.bsky.social · 28/09/2026
My latest post on Damage Labs ain't exactly rocket science. But it might involve a little. www.damagelabs.com/p/yeah-so-ro...
damagelabs.com
Yeah, so rockets...
The making of a high-power rocket flight
000
Scott Wasson @scottwasson.bsky.social · 17/09/2026
My latest at Damage Labs: Will Microsoft lose the consumer PC market? www.damagelabs.com/p/will-micro...
damagelabs.com
Will Microsoft lose the consumer PC market?
The PC is thriving. Windows is not.
000
Scott Wasson @scottwasson.bsky.social · 10/08/2026
My latest article at Damage Labs: How the AI hardware crunch will transform the PC www.damagelabs.com/p/the-person...
damagelabs.com
The personal computer's big-SoC future
The AI hardware crunch will transform the PC as we know it
000
Scott Wasson @scottwasson.bsky.social · 03/08/2026
Introducing Damage Labs! Some thoughts on 15 years of frame-time testing and the current state of things. www.damagelabs.com/p/inside-the...
damagelabs.com
Inside the second: where things stand today
15 years of progress in gaming performance—and some unfinished business
000
Scott Wasson @scottwasson.bsky.social · 29/07/2026
020
Scott Wasson @scottwasson.bsky.social · 24/07/2026
The Smurve80 looks like a Model M SSK, takes MX switches, has a curved plate like the old IBMs, and has a solenoid for the clack! Got mine and rebuilt it with Gateron Melodics.
151
Scott Wasson @scottwasson.bsky.social · 06/07/2026
Hey, folks. Just a quick note. I had my last day at MLCommons about a month ago. I'm now considering what's next in my career, and I'd be happy to talk with anyone about potential roles or ways we could collaborate. Let me know if you'd like to chat!
000
Scott Wasson @scottwasson.bsky.social · 04/07/2026
010
Scott Wasson @scottwasson.bsky.social · 16/06/2026
Found it.
000
Scott Wasson @scottwasson.bsky.social · 10/06/2026
This Assassin’s Creed II remaster is incredible.
100
Scott Wasson @scottwasson.bsky.social · 08/06/2026
Syryan Mont Cervin keyboard Carbon fiber plates Everglide Aqua King switches (205g0, Durock 63.5g slow springs, TX films) TX stabs MTNU WoB keycaps Createkeebs Venice Rialto keyboard PC plates Durock POM switches (205g0, CK 58g slow springs) Staebies NicePBT BoW with blue accents
061
Scott Wasson @scottwasson.bsky.social · 02/06/2026
Kiki’s Tented Alice keyboard FR4 plates Chosfox HMX Hanami Dango switches (205g0, Durock 63.5g slow springs, TX films) KKB WoB keycaps Black walnut wrist rest
050
Scott Wasson @scottwasson.bsky.social · 23/05/2026
Linworks/Syryan EM.8 Saffiano leather wrist rest inserts GMK Mictlán Noche Negra with Centinelas Prevail Epsilon switches (205g0, Durock 63.5g slow springs, TX films) TX stabilizers
082
Scott Wasson @scottwasson.bsky.social · 07/05/2026
Just arrived: Kiki’s Tented Alice by 27Studio.
000
Scott Wasson @scottwasson.bsky.social · 13/04/2026
351
Scott Wasson @scottwasson.bsky.social · 01/04/2026
MLPerf Inference 6.0 results are out! Lots of new goodness here, including a bunch of new benchmark tests (and models) and some very large clusters reaching new peak performance numbers.
010
Reposted by Scott Wasson
MLCommons @mlcommons.org · 17/03/2026
🎉 MLPerf Mobile is now on the iOS App Store! Run industry-standard ML benchmarks right from your iPhone. See how your device performs — download it now 📲 apple.co/4bvbfdI #MLPerf #AI #MachineLearning
121
Reposted by Scott Wasson
MLCommons @mlcommons.org · 08/01/2026
MLPerf Mobile app v5.0.4 is here! #MLPerf Mobile release now supports Samsung Exynos 2600 and Qualcomm's newest Snapdragon lineup, from flagship 8 Elite Gen 5 to mid-range 6 Gen 4. Meaning more comprehensive, apples-to-apples #AIperformance data across devices. play.google.com/store/apps/d...
play.google.com
MLPerf Mobile - Apps on Google Play
An AI benchmark for mobile devices
001
Scott Wasson @scottwasson.bsky.social · 17/11/2025
MLPerf Client v1.5 is out now. Updates include Windows ML support, GUI improvements, and experimental support for Linux and power-efficiency testing. We support a broad range of GPUs and NPUs across PCs, Macs, and iPad Pro. Free to download & run! mlcommons.org/2025/11/mlpe...
mlcommons.org
MLPerf Client v1.5 Advances AI PC Benchmarking with Windows ML Integration - MLCommons
MLPerf Client v1.5 introduces Windows ML support, expanded platform coverage (including iPad), and experimental power measurement, setting a new, comprehensive standard for benchmarking AI performance...
020
Scott Wasson @scottwasson.bsky.social · 20/09/2025
4GT Sabre Aluminum plate HMX Hades V2 switches (205g0, TX films, Durock 63.5g slow springs) TX AP screw-in stabs (205g0) GMK British Racing Green R2 Walnut wrist rest from Digital Carpentry
130
Reposted by Scott Wasson
MLCommons @mlcommons.org · 16/09/2025
We have released MLPerf Mobile v5.0.1 with support for MediaTek Dimensity 9400 SoCs and a few other improvements. The app is available for Android phones via the Google Play Store and the MLCommons GitHub repo. Let us know what you think! t.co/715ENpV6W4
t.co
https://play.google.com/store/apps/details?id=org.mlcommons.android.mlperfbench
001
Scott Wasson @scottwasson.bsky.social · 09/09/2025
MLPerf Inference v5.1 results are out today! New in this round: DeepSeek R1, Whisper V3, and Llama 3.1 8B tests. A bunch of new GPUs were tested, including AMD MI355X, Nvidia GB300, and Intel Arc Pro B60. mlcommons.org/2025/09/mlpe...
mlcommons.org
MLCommons Releases New MLPerf Inference v5.1 Benchmark Results - MLCommons
MLCommons Releases New MLPerf Inference v5.1 Benchmark ResultsNew results highlight AI industry’s latest technical advances
010
Reposted by Scott Wasson
Jeff Kampman @jkampman.bsky.social · 01/08/2025
MLPerf Client 1.0 is out! Happy to see an AI benchmark that supports as many platforms and as scalable a set of workloads as this one does, plus the new GUI is extremely handy.
tomshardware.com
MLPerf Client 1.0 AI benchmark released — new testing toolkit sports a GUI, covers more models and tasks, and supports more hardware acceleration paths
A useful new tool for navigating the highly fluid world of client AI performance
011
Scott Wasson @scottwasson.bsky.social · 31/07/2025
Another recent arrival: QK Alice Duo.
220
Scott Wasson @scottwasson.bsky.social · 30/07/2025
Just arrived: the 4GT Sabre.
100
Reposted by Scott Wasson
MLCommons @mlcommons.org · 30/07/2025
MLPerf Client v1.0 is out! 🎉 The new benchmark for LLMs on PCs and client systems is now available—featuring expanded model support, new workload scenarios, and broad hardware integration. Thank you to all submitters! #AMD, #Intel, @microsoft.com, #NVIDIA, #Qualcomm mlcommons.org/2025/07/mlpe...
mlcommons.org
MLCommons Releases MLPerf Client v1.0: A New Standard for AI PC and Client LLM Benchmarking - MLCommons
MLCommons Releases MLPerf Client v1.0 with Expanded Models, Prompts, and Hardware Support, Standardizing AI PC Performance.
002
Scott Wasson @scottwasson.bsky.social · 30/07/2025
MLPerf Client v1.0 is now available with a host of updates, including: -A beta GUI -3 new models -Large ~4K/8K prompts -GPU, NPU, & hybrid acceleration for all major PC hardware vendors -Early Windows ML acceleration on Intel -Win11 x64/Arm and macOS support
100
Scott Wasson @scottwasson.bsky.social · 10/07/2025
Happy to have MLPerf Mobile on the Play Store! Going forward, we're working to deliver proper hardware acceleration on new mobile devices as they hit the market, as well.
020
Scott Wasson @scottwasson.bsky.social · 17/05/2025
130
Reposted by Scott Wasson
MLCommons @mlcommons.org · 30/04/2025
As AI models grow, storage is key to #ML performance. MLCommons' @dkanter.bsky.social joins #Nutanix’s Tech Barometer podcast to explain why and how the #MLPerf #Storage #benchmark guides smarter #data #infrastructure for #AI. Listen: www.nutanix.com/theforecastb... #DataStorage #EnterpriseIT
nutanix.com
David Kanter MLPerf Measures AI Data Storage Performance
In this Tech Barometer podcast, MLCommons Co-founder David Kanter talks about creating the MLPerf benchmark to help enterprises understand AI workload performance of various data storage technologies.
012
Scott Wasson @scottwasson.bsky.social · 28/04/2025
Updated with support for Intel NPUs and more.
010
Scott Wasson @scottwasson.bsky.social · 18/04/2025
020
Scott Wasson @scottwasson.bsky.social · 09/04/2025
Ladies and gentlemen, the SplittyKitty. geekhack.org/index.php?to...
041
Scott Wasson @scottwasson.bsky.social · 03/04/2025
MLPerf Inference 5.0 results are out, with performance numbers for the latest AI accelerators, including AMD Instinct MI325X, Nvidia B200 and GB200, and Google TPU Trillium across a range of tests. mlcommons.org/2025/04/mlpe...
mlcommons.org
MLCommons Releases New MLPerf Inference v5.0 Benchmark Results - MLCommons
MLCommons' latest MLPerf Inference v5.0 results show Gen AI now the center of attention for performance engineering.
110
Scott Wasson @scottwasson.bsky.social · 31/03/2025
Here’s the launch of my level 2 high-power certification flight: an Apogee Zephyr flying on an Aerotech J270 engine. Went to ~4500 feet; recovered it in good shape to pass the cert test.
040
Scott Wasson @scottwasson.bsky.social · 21/03/2025
$3,600 keyboard. www.norbauer.co/pages/the-se...
norbauer.co
the Seneca
the Seneca a keyboard dream from the future Learn about buying from $3600 682 custom parts. Built with two human hands. The Seneca is my middle finger to the aesthetic homogeneity and economic over-op...
010
Scott Wasson @scottwasson.bsky.social · 02/03/2025
Yesterday’s launch.
050
Scott Wasson @scottwasson.bsky.social · 21/02/2025
240
Scott Wasson @scottwasson.bsky.social · 18/02/2025
Nature is amazing.
030
Scott Wasson @scottwasson.bsky.social · 26/01/2025
050
Scott Wasson @scottwasson.bsky.social · 25/01/2025
Yikes. www.reddit.com/r/Mechanical...
reddit.com
From the MechanicalKeyboards community on Reddit: Carnage
Explore this post and more from the MechanicalKeyboards community
020
Scott Wasson @scottwasson.bsky.social · 25/01/2025
Rama U80-A TKL with some cheap new translucent keycaps.
050
Scott Wasson @scottwasson.bsky.social · 24/01/2025
My current top MX clicky switches in no particular order: -Gateron Melodics -NovelKeys Cream Clickies -Zeal Clickiez -Kailh Box Whites
140
Scott Wasson @scottwasson.bsky.social · 24/01/2025
My current top linear MX switches in no particular order: -WS Pearls (and roller linear switches generally) -TTC Neptunes -Durock POMs -HMX Hades -Everglide Aqua Kings
140
Scott Wasson @scottwasson.bsky.social · 24/01/2025
Starting to suspect that local text summarization features aren't gonna fulfill the AI hype.
020
Scott Wasson @scottwasson.bsky.social · 24/01/2025
HSA profile beige on brown keycaps on the Keyclicks Split 65.
4253
Scott Wasson @scottwasson.bsky.social · 23/01/2025
Object-oriented juice.
021
Reposted by Scott Wasson
Nicholas A. Christakis @nachristakis.bsky.social · 15/01/2025
Incredible new work via @nature.com: deep learning was used to design proteins to bind snake venom. Proteins with remarkable thermal stability, high binding affinity, and near-atomic-level agreement with the computational models were made and proven by protecting mice from a lethal envenomation.
nature.com
De novo designed proteins neutralize lethal snake venom toxins - Nature
Deep learning methods have been used to design proteins that can neutralize the effects of three-finger toxins found in snake venom, which could lead to the development of safer and more accessible an...
2669
Scott Wasson @scottwasson.bsky.social · 13/01/2025
This header is no bueno.
271
Scott Wasson @scottwasson.bsky.social · 06/01/2025
And now we shovel.
000