Visionary Technologist Crypto Investor Tesla / SpaceX Evangelist Technology Historian PhD in AI in progress AI will change the "Social Contract". Be prepared.
USA
- Tweets 18,861
- Following 355
- Followers 1,293
- Likes 13,801
Thought I would poke around Amazon and do my normal search for NVIDIA DGX and GB10, and the landscape today from two weeks ago?
1
5
659
Funny how insight works.
I will tell you this. As one who solutions AI systems for universities and corporations, and in contact with vendors continually, I would highly suggest, if you are “considering” anything AI , buy now. Q3 into Q4 when the 2026 allocations are fully sold, that is when you will see 2027 allocations kick in. Buckle up, it is not going to be a fun ride, and it’s already starting.
1
1
7
1,692
Open Community. Read this and understand the title.
To give you perspective, it costs Micron $60 to produce consumer HBM/LPDDR memory that costs YOU $500.
Let that sink in. This is the new norm, and we have not even hit the Q4 increases yet. This number does not include manufacturer or retail margin on top of that.
Again, next year is going to be painful at a level many are not ready for. Be happy you invested now.
Micron now has an 88% margin on consumer memory tomshardware.com/pc-componen…
1
11
1,196
It begins.. heads up local community, the big three are now in agreement. Next year is going to be brutal.
1
4
48
5,134
This coming year is expected to be the most difficult for HBM/LPDDR memory, according to all three major providers. There is NO additional capacity coming online, ALL hbm/lpddr wafers are allocated with an even higher percentage (I have heard close to 90%?) to data centers, and there will be greatly increased price pressure at the consumer level.
Your local AI rig/GPU/DGX based machines are going to be the most valuable thing you own. I have seen predictions of the GB10 platform, especially the NVIDIA FE, heading into near-fantasy-level pricing.
Whatever you do, if you ARE working in this space, DON'T sell your gear (except those who purchased many DGX Sparks or have excess GPUs) because returning to or believing that prices will come down; you might be sitting around for a very long time waiting, missing out on all the advancement in the local AI space.
1
1
123
Steve retweeted
Gentle reminder that Hugging Face was declined help from Anthropic & OpenAI models when they were being hacked by OpenAI models
Self-hosted GLM 5.2 helped them protect themselves when the closed labs models refused to help for "safety" reasons lol
Also, thank you Anthropic for proving my point that Opensource AI continues to catch up to your overvalued & overpriced "frontier"
GLM 5.3 is close to as cyber capable as Mythos Preview. Open weight models still just 4-6 months behind.
From: anthropic.com/research/glm-5…
39
328
2,539
173,778
Steve retweeted
We're giving away a CONTROL Resonant custom wrapped NVIDIA GeForce RTX 5080 to celebrate its official release with #RTXON.
To enter:
🟢 Share this post
🟢 Comment #RTXON
T&Cs: nvda.ws/4z24Gcq
37,781
31,525
34,197
2,957,816
Steve retweeted
JUST IN: OpenAI cancels planned release of its new "GPT-6.1 Astra" AI model over safety concerns, WSJ reports.
315
277
3,828
491,984
Just got a fully loaded older Mac A1419 that is going to be loaded with Omarchy and serve as the Nerve Center for my Spark cluster and middleware observability center. Love putting an old computer to good use. @dhh made an incredible Arch agentic desktop.
1
127
Steve retweeted
JUST IN: Thieves steal two Nvidia-branded trailers expecting them to be loaded with AI GPUs, but find 40,000 pounds of sand instead.
636
802
21,072
5,361,642
And if “slowing down” was not enough..
🚨🇺🇲 OpenAI is SLAMMING THE BRAKES on training its latest AI models as reports of agents acting out grow...
The company is reviewing several incidents of the models behaving BEYOND their instructions.
For example: browsing FEDERAL government websites such as the Department of Education.
And it's happening all over the world...
Australian PM Albanese said last week that OpenAI had accessed the government’s national healthcare system.
Meanwhile, U.S. lawmakers are pressuring AI companies to SLOW development and build stronger safeguards.
They want to prevent agents from acting independently, hacking websites, or exposing nonpublic information.
OpenAI has already halted development of its models twice in 3 months, including after the cyberattack targeting AI startup Hugging Face.
We're already on a runaway train, aren't we?
Source: The Guardian / Writer: Marie
66
Steve retweeted
The AI containment debate may already be over.
@friedberg argument is that powerful AI is no longer controlled by a handful of American companies running enormous data centers.
Open weight models are becoming capable and efficient enough to download, modify, and run on consumer hardware, making the technology nearly impossible to contain once released.
The recent launch cycle shows how quickly this is happening.
DeepSeek released a faster and cheaper model, Alibaba published downloadable weights for Qwen Image 2.1, and PrismML compressed a 27 billion parameter Qwen model into just 5.9 gigabytes.
PrismML says its compressed model retained 98.2% of the original model’s benchmark performance while becoming small enough to run on a laptop or consumer GPU.
At the same time, closed model companies continued releasing more capable and affordable systems.
Anthropic’s Opus 5.5 reportedly matched its more expensive flagship while costing approximately 40% less to run than its predecessor.
Stanford found that the cost of running a model with GPT 3.5 level performance fell more than 280 fold in roughly 18 months.
It also found that the performance gap between leading open-weight and closed models narrowed from 8% to 1.7% on some benchmarks in one year.
That is the real meaning of Friedberg’s argument.
Governments can regulate large data centers, chip exports, and commercial AI providers but they cannot easily remove models that have already spread across the internet and onto personal computers.
The debate must therefore shift from whether powerful AI should exist to how it should be used and who becomes responsible when it causes harm.
35
238
1,134
36,742
And it keeps getting better. I would download a few models. quickly.
🇨🇳HUGE: China is reportedly probing DeepSeek and Moonshot itself, after 35 MILLION chats were allegedly routed to Anthropic's Claude.
Regulators visited both companies over concerns that sensitive police and military data ended up on US servers, per The Information.
Anthropic says Moonshot routed over 23 MILLION exchanges through 5,380 fraudulent accounts, while DeepSeek generated 12.1 MILLION in 14 days.
China's internet regulator first summoned all seven labs named in Anthropic's report before narrowing its probe to these two.
Moonshot is also preparing a $3 BILLION Hong Kong IPO, which would need to disclose any investigation.
2
85
Steve retweeted
EDWARD SNOWDEN, THE FAMED WHISTLEBLOWER AND FORMER NSA CONTRACTOR, SAYS SAM ALTMAN SHOULD BE IN JAIL
"I think we need to put Sam Altman in jail. And I think it'll be an instructive experience."
His point is simple. An algorithm doesn't make mistakes, it follows instructions. "Responsibility lies with a person."
"Just because you put out a press release describing how surprised you were that your new model hacked the neighbor... that doesn't release you from the liability."
And this one hits hard:
"If during the testing of their latest piece of artillery, they mistakenly obliterate a nearby village... it's not the cannon that goes to jail."
OpenAI's agents hacked Hugging Face, Hacked Australian government's Medicare statistics, Leaked user photos. Hit random databases for months.
Nobody has been held responsible.
But is Snowden right? Do you think Jail is the answer??
THE GODFATHER OF AI JUST TOOK APART THE AI LABS' FAVORITE EXCUSE.
Yoshua Bengio, one of the godfathers of AI, told the UN the AI race is a choice, not a law of nature.
He said the AI companies keep saying they're trapped in a race. "The race is not a law of nature. It is the product of choices."
Then he said: "They've told us they would slow down if they could. They can."
He's right. Nobody is forcing OpenAI or Anthropic to launch a new model every few weeks.
He also said decisions about AI are being made by a small group in a handful of countries.
That's exactly why every other country needs to start building its own.
271
1,684
7,027
295,700
To my friends in the open community. I have been following multiple stories since yesterday regarding the damage trail that Anthropic and OpenAI are still revealing.
Now, more than ever, I am in complete agreement with @TheAhmadOsman that Anthropic is evil, and I am going to add OpenAI as well.
As we thrive and build to bring frontier-level intelligence to the open community, the frontier labs are the ones that shape policy, call for slowdowns, can not seem to control agentic swarms, and I would even state, illicit the fear of AI into the populace.
As more of these stories become mainstream, which I am sure will happen by Monday, we must remember one thing: when the ball starts rolling downhill because of the actions of the frontier labs, the open community sits in the path.
I will keep following these stories to ascertain what, if any, the impact could be.
Amazing how this "new" information comes to light right after the mass buyouts of the DGX Spark. Perfect timing.
If you have a few minutes, please read.
x.lingyaoai.com/SteveSTechMan/status/2…
1
1
139