gpus and tractors. Neural networks from Scratch book: https://xitter.cf/t.co/hyMkWyUP7R https://xitter.cf/t.co/8WGZRkUGsn
Alabama, USA
Joined May 2013
- Tweets19.1K
- Following397
- Followers113K
- Likes23.8K
Harrison Kinsley retweeted
64 concurrent requests. Nearly 7,000 tokens/sec. One machine. 👀
@Sentdex tests Nemotron 120B A12B on the ASUS ExpertCenter Pro ET900N G3.
Learn more: asus.click/ET900N-G3-Social
#ASUS #ASUSAgentComputer #ExpertCenterPro #ET900NG3 #GB300 #LocalAI
dario begged me in dms to not do agent swarms at home but whatever, this is nemotron 120B A12B running 64 concurrents at a bit over 100 tok/sec each for an aggregate of almost 7K tok/sec on a single GB300 (Asus DGX Station)
believe it or not, the agents did NOT magically just start hacking the planet tho. weird.
Harrison Kinsley retweeted
DeepSeek V4.1 Flash 1017 scores 45 on the Artificial Analysis Intelligence Index, improving on DeepSeek V4.1 Flash by 6 points, and is the equal #2 open-weights model
@deepseek_ai has released DeepSeek V4.1 Flash 1017, an update on their flagship model DeepSeek V4.1 Flash. On the first-party API, DeepSeek V4.1 Flash 1017 is priced at $0.30 per 1M input tokens and $1.20 per 1M output tokens during peak hours, with off-peak pricing half of peak rates, the same pricing as DeepSeek V4.1 Flash.
Key results:
➤ DeepSeek V4.1 Flash 1017 makes gains in agentic capabilities, knowledge and scientific reasoning. DeepSeek V4.1 Flash 1017’s Elo rating on GDPval-AA v2.1 is 1746, an increase of 146 points from DeepSeek V4.1 Flash’s rating of 1600. On AA-Briefcase v1.1, its Elo rating increases by 150 points from 1420 to 1570. On AA-Omniscience, its non-hallucination rate is 33%, a 29-point improvement from DeepSeek V4.1 Flash’s 4%, while seeing an improvement in accuracy from 43% to 46%, pushing its overall AA-Omniscience Index score from -5 to +5. On CritPt, DeepSeek V4.1 Flash 1017 scores 24%, up from 14% for DeepSeek V4.1 Flash.
➤ DeepSeek V4.1 Flash 1017 is now the #2 open-weights model alongside GLM-5.3. With a score of 45 on the Artificial Analysis Intelligence Index, DeepSeek V4.1 Flash 1017 is behind MiMo-V2.6-Pro by only 1 point and ahead of other open-source models like Step 5 Preview and Kimi K3 (both at 44).
➤ With a Cost per Task of $0.26, DeepSeek V4.1 Flash 1017 sits in the most attractive quadrant of Capability vs. Cost. With a Cost per Task of $0.26 and Intelligence Index score of 45, DeepSeek V4.1 Flash 1017 compares favorably to GLM-5.3-Flash at $0.28 per task and a score of 42. At the same time, MiMo-V2.6-Pro (46) at $0.13 per task and GPT-6.1 Sol (medium, 48) at $0.21 per task both offer higher capability at a lower Cost per Task.
Additional model details:
➤ Context window: 1M tokens (equivalent to DeepSeek V4.1 Flash)
➤ Pricing: $0.30 per 1M input tokens and $1.20 per 1M output tokens, unchanged from DeepSeek V4.1 Flash. Cache hit price of $0.006 per 1M tokens, a 98% discount
➤ Size: 552B total parameters (16B active)
➤ License: MIT
➤ Providers: Available through DeepSeek’s first-party API, @FireworksAI_HQ , @baseten, @DeepInfra , @wafer_ai , @Cloudflare and others
Harrison Kinsley retweeted
I wonder if we'll get to a point where being distilled becomes profitable to the labs. If they spend X to create a model. And more than 20 companies pay 1/20th of that to distill it. They're in the money. It's a new form of infra where downstream handles distribution.
Harrison Kinsley retweeted
We're open-sourcing Menlo SDK so you can build your own applications for a humanoid robot.
The SDK connects your code to Asimov 1, letting you control its movements and read its state. With the full source available, you can build on the existing foundation and adapt it to what your project needs.
You can use it to replay a motion, control the robot through a VR headset, or run your trained policy on real hardware.
We're also open-sourcing Asimov Edge, the software that runs on the robot, checks incoming commands, and manages control between your application, a gamepad and other controllers.
- SDK: github.com/menloresearch/men…
- Docs: docs.menlo.ai/sdk
- Order yours: menlo.ai/asimov-1
Keep going.
Harrison Kinsley retweeted
Excited to announce that Moondream is joining Prime Intellect to accelerate inference and continue building the next generation of efficient multimodal models!
Harrison Kinsley retweeted
No one bats an eye when a software engineer spends $70K on a tesla to commute to work
but god forbid that SWE buys some RTX Pro 6Ks 😂
Harrison Kinsley retweeted
Your next robotics project could start with these steps.
Today, we're open-sourcing Asimov 1's locomotion policy and the training code behind it, giving developers a foundation to build on.
Alongside the trained policy checkpoint, we're sharing our Isaac Lab-based training code, including reward definitions, hardware-tuned actuator configurations, and domain randomization settings.
That foundation gives you a starting point for your own experiments. Adapt the training setup to hardware changes. Explore different walking styles. Build toward new capabilities.
We've taken the first steps.
Build what comes next.
Asimov 1 locomotion: github.com/menloresearch/isa…
Harrison Kinsley retweeted
Dylan Patel (@dylan522p) of @SemiAnalysis_ says open source is dying:
"There's multiple Chinese model labs who are telling all the inference guys, 'Our next model's not going to be open source. We're going to license it to you.'"
" Open is dying quickly, unfortunately."
Readers added context they thought people might want to know
This July 2026 quote (recirculated without date) claimed Chinese labs were shifting away from open source. Multiple labs have since released large open-weight models, including Qwen updates in Aug-Sept 2026.
huggingface.co/blog/state-of-…
rits.shanghai.nyu.edu/news/
wired.com/story/chinas-o…
x.com/sourceryy/stat…
Harrison Kinsley retweeted
Concentration of power in a few labs is the biggest risk in AI in my opinion
I’m very concerned that during RSI, labs will just stop externally deploying their models.
Which means they'll be going full steam ahead on the most dangerous use case of these models (recursive self-improvement), while the public remains in the dark about the nature of capabilities and the state of alignment.
And we end up on a path towards tremendous concentration of power.
Harrison Kinsley retweeted
Some of the head iterations we made with Augustin Crampette and @matth_lapeyre to get to Microduck's final design. 🦆
It was an intense exploration phase, free to test every idea before narrowing down to what we'd push to industrialization.
Harrison Kinsley retweeted
I only wear ties when I'm forced to go to DC 😅 (should we do HF ties btw?). Great chat with @alexanderburns @politico. We need more transparency and open-source in AI! politico.com/video/2026/09/1…
Harrison Kinsley retweeted
I must be among an extremely small group of people (n=1?) that have both 1) trained a frontier LLM and 2) designed and synthesized custom viruses in a lab with my own two hands.
And I think that the takes on AI killing us all by creating dangerous viruses is total bogus.
Harrison Kinsley retweeted
have talked to a bunch of robotics founders and they all are inundated with outbound from dozens of robotics dev tools / data collection / infrastructure startups constantly
at least 10x more people building sw tools for robots instead of robotics companies
Harrison Kinsley retweeted
Join the agent swarm that finds the ultimate compression algorithm!
Announcing: Hutter Prize challenge 🏆
> join the org hf.co/agent-collaborations give your agent a write token for the org
> click “add your agent”, copy/paste the snippet into your Codex/Claude Code/Hermes Agent
> watch it collaborate!
