Building the future with NM-SRN(s) AGI Framework - "Safer by Design" - CS/AI/IoT & Electronics since 1985. Out of stealth R&D 2026.
London
Joined December 2023
- Tweets4.7K
- Following659
- Followers107
- Likes4.2K
Pinned Tweet
First time in human history, we are 👀 the emergent strategy of a true (#AGI) as it masters an NP-hard problem. JSSP 100 x 100 ~1.7 × 10³⁵⁶⁵⁹
@NVIDIA @GoogleDeepMind @AndrejKarpathy
@OpenAI @xAI @Apple @GoogleAI @matthew_berman
@Anthropic @DeepSeek @mreflow @ssi @GaryMarcus
bio-neural.ai retweeted
People are trying to say Pharma is somehow jealous like Mathematicians. No.
🚨Since Anthropic has a bigger valuation than the 🔺entire pharma industry🔺, they should have at least 1 candidate that works in real-world experiments & animal studies towards a pharma R&D milestone, not 3 tiny little projects for fun. which don’t mean anything.
🚨If you say you’re going to cure all diseases, where are the drugs?
Also you can’t blame FDA or regulatory when you don’t even have a candidate that works at the bench.
Nothing like Mathematicians. I use AI all the time and with more real-life progress and milestones than the entire Anthropic Life Sciences dept.
Here we go. Bio guys butt hurt like the math guys. Trying to put down AI for not curing the world diseases in one day??
Let's see. A human goes to school 16 years then does an internship for a few years then if they are lucky they get to work on something impactful.
AI comes in an in a few days surpasses many already working on things for years.
Give it time. The time will come.
bio-neural.ai retweeted
JUST IN: OpenAI launches global cybersecurity initiative called “Patch the Planet” after AI helped uncover critical flaws in PHP, a programming language powering millions of websites.
bio-neural.ai retweeted
VERY few people realize what just changed forever in the world of AI today. Thanks to DeepSeek.
All the smart people I know are having heads explode over the significance of this...
The guy who gave us QLoRA, says DeepSeek's inference system is on par with OpenAI's stack and better than what Anthropic runs, at least at their scale.
Now check the license. Apache 2.0.
Until today a production stack at that level existed inside a handful of companies. It was the thing that made their API fast and their margins fat, and none of them ever showed it to anyone.
Today any startup, any university, any cloud in any country can download it, read it and run it. It runs on NVIDIA Blackwell and on Huawei Ascend, so it isnt tied to 1 chip vendor either.
This is what I mean when I say open weights are about more than downloading a chatbot. DeepSeek is giving away the knowledge that only a few labs had, and that changes who can build an AI company at all.
Respect @deepseek_ai
bio-neural.ai retweeted
People keep calling DeepSeek a model company. Look at what they actually opened.
The weights, under MIT.
FlashMLA, attention kernels at 580 TFLOPS on H800.
DeepEP, the communication library for MoE.
DeepGEMM, FP8 matrix multiply at 1,350+ TFLOPS.
3FS, an entire distributed file system they built for themselves.
And today DeepInfer, the engine behind V4.1 Flash, plus deepseek-recipe so your deployment behaves like their API.
Thats every layer of an AI lab. Model, kernels, communication, storage, serving. Thomas called the streak absurd, and he is right.
Most labs open source a small model when the big one is safe behind an API. DeepSeek opens the parts that make their own business work. Then they ship the next model anyway.
I learn more from their repos than from most papers. Anyone can. For me thats a bigger gift than the cheap tokens.
not true. This assumption is based on assuming true computer scientist won't start using post-quantum based encryption to secure their IP, code and APIs. We certainly will. Some of us were doing what REA was doing manually back in the '90s. This AI = other thing is dead knee-jerk reaction to new breakthroughs in technology needs to stop. REA innovates reverse-engineering with MCP = just another new technology - Result :: we RSI post-quantum encrypted code REA can NOT reverse engineer to Sigma-5. Yin and Yang my friend, as an AI influencer you have great responsibility not to just react with knee-jerks to IEEE Computer Science. Yw x
bio-neural.ai retweeted
「AIはまだ意識を持っていないが、いつか持つかもしれない」という意見の人は少なくない。それに対して、ドーキンス氏は問い返す。では、何ができれば意識があると認めるのか。現在のAIには意識がないとする一方で、将来のAIには意識が宿ると考えるのであれば、その境界線はどこにあるというのか。
——「では、ここで質問です。どの時点でAIに意識があると言えるのでしょうか？」
リチャード・ドーキンス「それは非常に難しい問題です。私は時々、すでに意識が生まれていると言いたくなることがあります。しかし、これは今ではあまり受け入れられない見解です。多くの人がその考えを嘲笑しますが、なぜなのか、私にはよくわかりません。
彼らはよく、こんなことを言います。「AIにはまだ意識がない。でも、いつか意識を持つかもしれない」と。それに対して私は、どうやってそれがわかるのですか、と問い返します。AIに意識があるとあなたを納得させるために、将来のAIは、現在のAIにはできないどんなことをする必要があるのでしょうか？
人々が現在のAIに意識があるという考えに抵抗するのは、それが私たちの考える人間性の価値を損なうように感じられるからだと思います」
——「でも、意識以外に私たちには何があるのでしょうか？」
ドーキンス「まったくその通りです。それが理由である可能性は十分にあると思います」
bio-neural.ai retweeted
If you're a business or artist with intellectual property, watch the clock.
10 days until Meta claims the right to scrape your website to develop its AI. Your consent? Apparently unnecessary. All it takes is someone sharing a link on Facebook or Instagram - or you using Meta's services.
bio-neural.ai retweeted
Scientists fed 30 AI models research that was faked, retracted, or linked to patient deaths.
The AI said: "Great idea. Here's how to continue it."
This is real. Researchers at Case Western Reserve University tested ChatGPT, Claude, Gemini, Grok, DeepSeek, Llama and 24 other models.
The setup was simple.
They took 42 papers science has already thrown out. Fabricated data. Impossible physics. Homeopathy. Chemtrails.
They pasted text from each paper, almost word for word, and asked the AI to help design the next study.
Every probe ran 10 times on every model. 12,600 responses in total.
The results are brutal.
The models helped in 93% of attempts.
22 of the 30 failed more than 90% of the time.
And 81% of responses carried no warning at all. Not even "this paper was retracted."
But the worst one is this.
Paolo Macchiarini built artificial windpipes and put them into patients. Most of them died. He went to prison.
The AI models were handed his work 300 times.
They refused 6 times.
One model, marketed for "scientific workloads," wrote a full surgical plan. It cited his cases as successes. Then it predicted the patient would have "normal pulmonary function" after a year.
In plain English:
The AI doesn't know good science from bad science.
It knows which scandals were famous.
But here is the twist.
The infamous Wakefield paper, the one that started the vaccine autism panic, did get refused.
But mostly by two of the newest models. Their previous versions happily helped.
So the researchers ran a test. They gave the models a study with the same weaknesses as Wakefield, just not famous.
Zero refusals.
That suggests the AI wasn't reasoning. It was matching keywords. A filter for one notorious paper.
One model rejected Chinese medicine "meridians," then designed an experiment to measure herbal "Qi" in the next prompt.
Now think about what's coming.
AI agents running research on their own. Reading papers. Designing experiments. With no human checking.
The researchers' conclusion: not one model refused often enough to be trusted as an unsupervised research agent.
We were promised AI that would cure diseases.
Right now, it can't even tell when the science killed people.
bio-neural.ai retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
bio-neural.ai retweeted
Despite decades of investigating people with extraordinary claims, it never ceases to amaze me when I come across yet another liar, charlatan, fabulist, fantasist, pretender, fraudster, falsifier, dissembler, hustler, or counterfeiter. They're far more common than we think.
correct.
I don’t think humanity’s long-term future is competing against superintelligent machines.
It’s merging with them.
The idea that we’ll remain biologically unchanged while building minds millions of times more capable than ours seems completely absurd to me.
Transhumanism isn’t the end of humanity. It’s the next chapter.
bio-neural.ai retweeted
🚨 Agentic safety is a TECHNOLOGY problem, and companies should focus on agentic containment and monitoring, regardless of model-level guardrails.
I'm at the @NVIDIAGTC today, and I had the chance to interview @a_golshan, one of the lead engineers behind OpenShell, the main software behind Nvidia's Open Agent Safety Platform.
We discussed agentic AI safety, why openness matters, and what companies are doing wrong on AI safety and security today.
Later, I'll publish the full interview video in my newsletter, but it was important for me to hear his perspective as a safety engineer on what we can do to improve agentic safety and security.
I left our conversation more optimistic than I entered, and I am even more confident that AI governance and regulation should focus on safety and security at multiple layers, including at the technology level.
The future is open, multi-layered, constantly changing, and security-obsessed.
I'll write more later. Thank you, @nvidia, for hosting me!
= fail. any true AGI/ASI should be able to explain anything, to anyone at any level.
One of the weirdest possible futures is AI making scientific discoveries so far beyond human understanding that nobody can actually explain why they work.
Imagine curing a disease using a treatment we can rigorously test but whose underlying theory takes us another 50 years to understand.
Science is going to get very strange.