
CEO AIFlow Labs https://xitter.twitterapi.xyz/t.co/L27ts7lVwj CTO @datainetwork AI Advisor
Phuket
Joined October 2023
- Tweets16.3K
- Following5.8K
- Followers4.4K
- Likes67.8K
Container Desktop v0.1.0 is out: an open-source, native Mac window for Apple's container runtime. Run, inspect and manage Linux containers in a layout you already know. MIT licensed.
Article: xitter.twitterapi.xyz/AIFlow_ML/status/21121…
Repo: github.com/AIFlow-Labs-Limit…
#macOS #opensource
2
11
675
Jensen join the new Fab project will be the best thing to happen. Ah also Nvidia at 110 so they learn to LOI properly
53
Dear MacKenzie Scott Igor here.
If you have 8M hanging out of the Hermes bag please send to me to buy GPUs.
Jeff Bezos's divorce cost him 4% of Amazon. His ex-wife MacKenzie Scott walked away with $38.3 billion in stock and then started giving it away.
They were married for 25 years and have four children. She was there when Amazon was just getting started.
The divorce was finalized in 2019. Her share was about 4% of the company, and it's still called the most expensive divorce ever.
Then the pandemic sent Amazon stock up. For a while her fortune passed $66 billion and she was the richest woman in the world.
After the divorce she signed the Giving Pledge, a promise to give away at least half of her wealth.
By 2025 she had given more than $19 billion to over 2,000 organizations, with no strings on how they spend it.
In 2022 Bezos told CNN he plans to give away most of his money. He still hasn't signed the Giving Pledge.
93
DS5 in pretraining on Huawei !!
Wild rumor coming out of China this morning on DeepSeek V5:
DeepSeek is reportedly already pretraining V5 as a ~2 Trillion parameter ultra-sparse MoE model.
And instead of training on NVIDIA GPUs, they are reportedly running the V5 pretraining run on Huawei's new Ascend 950 / 950DT chips with a team of Huawei kernel engineers physically working inside DeepSeek's offices to keep the cluster humming.
If DeepSeek pulls off a 2-trillion-parameter frontier run natively on Ascend 950 silicon, that changes the entire AI hardware story.
28
Uncle is under fire.
The mud machine that we know is pushing to oust him for what ? An affair ? A spy ?
No kids. They want to nationalize Nvidia.
Jensen is one of us.
56
Yeah !!!!!!!!! DC will get bank on his feet
Woke Producer James Gunn Fired From DC After Box Office Failures zerohedge.com/political/woke…
49
So Malesya came up with a visa to build there.
Pay 12k
No tax break
No grants
No help
But you can relocate your fam.
Im pretty sure they will attract a lot of people running from something and not the best of the best.
52
I cannot wait to see it in Shenzhen !
#MaixStation Dock PCB come back~
Still waiting for 14-layer SOM PCB...
sipeed.com/maixstation
1
78
Can we have one desktop DGX killer with 256gb ram ?
🚨Huawei's Kirin 9050 series can run a 30B MoE model on smartphones, fully offline.
Huawei detailed the Kirin 9050's NPU today: a Da Vinci core in the first LogicFolding chip.
NPU: +140% performance, -66% power vs last gen.
It packs a Transformer engine, a low-power perception engine, and Huawei's CANN architecture.
It also supports:
• 6B image models
• 7B Dense language models
Text generation is 180% faster, with 29% less power consumption.
It also enables offline AI features like eye-tracking page turns, XMage Smart Capture 2.0, and screen-off maps.
56
Igor Lessio - Robots/acc - AIFlow Labs retweeted
CHINA HAS JUST MADE LIGHT TRAVEL FASTER THROUGH THE INTERNET! And this could be a game changer for AI.
While much of the world is focused on who can build the fastest AI chips, Chinese researchers have been tackling another major challenge: how to move enormous amounts of data between computing centres at incredible speeds.
Researchers from China Telecom, ZTE and Yangtze Optical Fibre and Cable have successfully tested a new type of optical fibre over a distance of more than 1,000 kilometres, reducing transmission delays by around 30%.
Now, here's the clever part.
Traditional fibre-optic cables transmit light through solid glass, which actually slows the light down. This new technology uses something called hollow-core fibre, where light travels mainly through air, allowing it to move much closer to its maximum possible speed.
The results? An astonishing 64 terabits per second in each direction, simultaneously, with signals travelling 1,000 kilometres in just 3.43 milliseconds.
To put that into perspective, that's enough bandwidth to transmit around 160 full-length 4K movies every second, in each direction!
But what's really interesting is what this means for artificial intelligence.
As AI models become increasingly sophisticated, massive data centres need to communicate with each other, often across enormous distances. Reducing those communication delays could make distributed AI computing considerably more efficient.
And this fits perfectly with China's strategy of building computing infrastructure across the country, taking advantage of cheaper energy in western regions while serving the major technology centres in the east.
Of course, this was a controlled transmission trial rather than a fully deployed 1,000-kilometre commercial network. But it demonstrates just how far the technology has progressed.
What I find fascinating is that China isn't simply concentrating on developing more powerful processors. It's investing across the entire technology ecosystem, from semiconductors and AI models to energy infrastructure and now the networks connecting everything together.
This is something I've said many times: don't underestimate China.
While everyone is watching the AI chip race, China is quietly working on the infrastructure that could help power the next generation of artificial intelligence.
And sometimes, the biggest technological breakthroughs aren't the ones you can see.
WTF when im in Thai Events when im in HK only sales hahahahaha
I want more robots
Best robotics meetup in HK is happening again tomorrow with great presentations like Adam’s. RSVP: luma.com/d8y1fsrc
58
Working with Fable 5.2 is so good guys.
I had ZERO issues till the preview. Only problem is that is eat the whole tokens available in 15 min.
I really like how it can handle very difficult tasks.
Good job @AnthropicAI
1
95
Any humanoid that we are adding into the produciton to test should not have legs but a solid base and wheels.
Like the @DexForceAI one for example.
47
Check out my #NVDA analysis on @TradingView: tradingview.com/x/kuzo8dvq/
1M not oversold and still on the main trend.
2
107
Who is building Monster Duck ?
I want one that actuallhy talk and have a laser 3000W that cut please :D
47
Imagine when some bad country will use it against us....
Anyway spectacular !!
Norway now has more than 2,000 underwater AI laser robots shooting lice off its salmon
they're made by a company called Stingray, and they already guard more than 50 million salmon and trout 24/7
the reason is sea lice, tiny parasites that latch onto salmon and feed on them
as few as 11 can kill a small fish, and one pen can hold 150,000 salmon
the usual fix is pumping the fish out and blasting them with warm water or water jets to knock the lice off
but the handling is so rough that Norway's fish health report ranks it the top cause of death in full-grown farmed salmon
so instead, Stingray puts a robot inside the pen that hunts lice one at a time, all day:
> two cameras watch every salmon that swims past
> AI checks two frames to make sure it's really a louse
> then a green laser fires from up to 2 meters away
> the salmon's scales reflect the beam, so the fish swims off fine and the louse is dead
we now live in a world where AI lasers do pest control on your dinner lmao
78





















