Bart S.

Hacking and building Things with AI • 15+ years engineering

OpenThings

[Bart S. @bartdarebel]

Doomsday (opinionated) backup:

- DeepSeek-V4-Pro (1.6T) https://t.co/LiRGSrYVvl
- Kimi K2.6 (1T) https://t.co/JKAXUrJraW
- GLM-5.2 (753B) https://t.co/VkUtehZFHy
- MiniMax-M2.7 (230B) https://t.co/sn3xiguK8h
- Qwen3.5-397B-A17B https://t.co/UUNGQercDa
- Qwen3.6-27B https://t.co/H61haM8SyA
- DeepSeek-V4-Flash (284B) https://t.co/nOGE6gkZvX
- Gemma 4 31B https://t.co/d1qmNP9NxY
- Qwen3.6-35B-A3B https://t.co/nWEkR5pjyw
- Gemma 4 26B-A4B https://t.co/bTdNczUTLE
- GLM-4.5-Air (106B) https://t.co/9hwfPXSO0V

[Jukan @jukan05]

CHINA CONSIDERS RESTRICTING OVERSEAS ACCESS TO CUTTING-EDGE AI MODELS

China’s Ministry of Commerce has led meetings over the past month with major AI companies, including Alibaba, ByteDance, and https://t.co/YDe0KRldDB, to discuss measures that would restrict overseas access to cutting-edge AI models, including models that have not yet been released.

The discussions reportedly include not only closed-source models but also open-weight models. However, the scope of application is still under debate, and the rules may ultimately apply only to future frontier models.

Officials have also discussed designating the leakage or theft of proprietary AI technologies as a national security crime, with stronger penalties, as well as restricting the types of foreign capital that can invest in Chinese AI startups.

The backdrop is the U.S. move to strengthen export controls on AI models, along with national security concerns over cutting-edge models that could possess advanced cyberattack capabilities.

Chinese authorities are reportedly concerned that advanced U.S. cybersecurity AI models could be used to exploit vulnerabilities in Chinese software.

Since the beginning of this year, China has continued to tighten measures to prevent AI technology from being transferred overseas. Authorities have investigated whether Chinese AI startups that relocated abroad violated export control laws, while also strengthening oversight of overseas transactions involving Chinese investors, technology, data, and national security concerns.

Future regulations could take the form of a tiered framework based on technological capability. Basic open-source AI models may be managed through a filing system, high-performance models may be subject to security reviews, and the most sensitive frontier models may be banned from public release or restricted to use within China.

[Bart S. @bartdarebel]

@marclou Uuu nice quote Marc. Hyrox prep?

[Bart S. @bartdarebel]

@theroborourke Don't skip them legs!

[Bart S. @bartdarebel]

Making Internet of Things (IoT) digital twins doesn't have to be boring.

Thanks @threejs for everything you're doing.

#iot

[Bart S. @bartdarebel]

3D diagrams somehow are more pleasing IMO

[Bart S. @bartdarebel]

@0xSero @yacineMTB 3 > 2

[Bart S. @bartdarebel]

@theojaffee

[Bart S. @bartdarebel]

@HelloVyom

[Bart S. @bartdarebel]

@rethink_hub Hardly to disagree

[Bart S. @bartdarebel]

@ProtonDrive Europe FTW!

[Bart S. @bartdarebel]

I'm looking for people who are interested in Over-The-Air updates (OTA) on their connected devices. Either bare metal or Linux.

I'd like to understand the journey they went through to build/code OTA on their devices.

#ota #update #overtheair #iot #connecteddevices

[Bart S. @bartdarebel]

@boardyai How does that differ from a Forward Deployed Engineer (FDE)?

[Bart S. @bartdarebel]

@DavidOndrej1 No cmux anymore?

[Bart S. @bartdarebel]

I just reserved my Cloudflare Wallet tag: https://t.co/NftMYzo8FU.

Reserve yours now at https://t.co/GOBpF6JhDi

[Bart S. @bartdarebel]

Do you know any Indie hackers that are living from embedded or hardware?

[Bart S. @bartdarebel]

@DanielLockyer Amen!

[Bart S. @bartdarebel]

@levelsio @DanielLockyer I think I might follow @DanielLockyer 🤔

[Bart S. @bartdarebel]

This should be fun.

[Liquid AI @liquidai]

Today we release LFM2.5-2.6B, an agentic model that runs entirely on-device. It plans, calls tools, and works through multi-step tasks on phones, laptops, PCs, and robots. Data never leaves the device, and the marginal cost of each run is essentially zero.

> Pre-trained on ~34T tokens
> LFM2.5 flagship hybrid architecture
> Context length: 128K
> Vocab size: 128K
> balanced intelligence per watt
> customizable on a single GPU for any specialized task
> LFM2 open-weight license

Comparable or better scores compared to models up to nearly 4x its size:
> ToolSandbox 77.83, ahead of Qwen3.5-9B at 76.44
> Multi-IF 80.07, ahead of Gemma-4-E4B-it at 77.35
> IFStruct 85.49, ahead of Qwen3.5-9B at 78.50

🧵

[Bart S. @bartdarebel]

@Hesamation Google is investing in Jeff’s company.

[Bart S. @bartdarebel]

@demishassabis I was wondering when such a wonderful chess move would happen. Congrats @demishassabis to you and the Team.
Looking forward to cross paths someday in the future.

[Bart S. @bartdarebel]

@sundarpichai @GoogleDeepMind @demishassabis @IsomorphicLabs Congrats to you and the Team @sundarpichai, that’s a very wise and strategic decision.