Nothing here is financial advise
Canada
Joined May 2009
- Tweets74K
- Following4K
- Followers3.1K
- Likes69.7K
Kirill 🔰 MBA retweeted
The biggest unreported story from yesterday: Carney snuck into the budget a massive immigration increase with the LARGEST NUMBER OF NEW PERMANENT RESIDENTS EVER next year.
Nearly 200,000 more than Trudeau’s highest-ever number year. When we have a shortage of jobs, homes, and health care.
Now we know why Carney snuck his immigration plan into a budget when everyone was distracted by the midterms and right before Parliament shuts for 10 days.
Unacceptable. This needs a real debate.
Kirill 🔰 MBA retweeted
🚨 ATTENTION: Most People Never Touch the Layer of Mind that Already Has the Answers 🚨
The mechanism that activates it is straightforward once you see it.
After that, ordinary thinking shows up as a habit you were maintaining, not the only way the mind can run.
The difference is binary...
Use the mechanism and the constant negotiation with your limits stops.
Leave it unused and the same narrow thoughts keep circulating until the end.
Once the mechanism is clear, the previous mode becomes visible as a choice rather than a fixed condition.
That visibility is the actual separation between staying ordinary and operating from the part that already holds the solutions.
Now that you have this knowledge, what will you do with it?
Let me know what You think,
and SHARE THIS so that
others may too.
And if
You're not
Already Following
@TrueOnX...
What
the Heck
Are You Doing?!?
🚨 DAVID WILCOCK’S FINAL MESSAGE just dropped… and it might be the most important thing he ever said.
In one of his last recordings before his passing, the legendary researcher breaks down the REAL science and ancient secrets of the Pineal Gland; your body’s built-in “third eye,” piezoelectric crystal antenna to higher consciousness.
They’ve been calcifying it for decades with fluoride, junk food, and toxins… because an awakened pineal gland = awakened humanity.
Watch until the end. Then go decalcify yours. David Wilcock.
Your work awakened millions. The pineal activation continues.
Let me know your thoughts below and SHARE THIS so that others may too.
Kirill 🔰 MBA retweeted
OpenAI 把整套 Codex Harness 开源了！
除了 CLI，还有 Skill 和运行环境，想学习 Agent 构建的朋友们有福啦！
github.com/openai/codex
Kirill 🔰 MBA retweeted
Your local LLM can now take control of a browser and complete tasks autonomously via Unsloth!
Qwen3.8-27B successfully controlled the browser to complete tasks like chess, painting, with no human intervention.
GitHub: github.com/unslothai/unsloth
Guide: unsloth.ai/docs/basics/brows…
Kirill 🔰 MBA retweeted
LFM3-1.7B just got an uncensored version, and it's small enough to run on almost anything.
• Quants come in at 1.1 GB and 1.6 GB
• Refusals dropped from 61.7% to 0%
• Claims no accuracy loss
THE SAFETY LAYER LASTED LESS THAN A WEEK. CAN OPEN MODELS EVER STAY LOCKED DOWN?
Open Weights: huggingface.co/AtomicChat/LF…
Kirill 🔰 MBA retweeted
holy f*ck, Karpathy put an entire AI degree on GitHub for free
7 repos, 306,741 stars, and in the right order they take you from backprop to agents that run research overnight
each repo is the next layer of the one before
1. micrograd -> how a neural net actually learns
backprop in about 100 lines of Python
↳ github.com/karpathy/microgra…
2. nn-zero-to-hero -> the video course
his lecture notebooks, from one neuron to GPT
↳ github.com/karpathy/nn-zero-…
3. minbpe -> how text turns into tokens
the tokenizer every LLM starts with, in clean minimal code
↳ github.com/karpathy/minbpe
4. nanoGPT -> your own GPT
the simplest repo to train and fine-tune one
↳ github.com/karpathy/nanoGPT
5. llm.c -> the same training with nothing hidden
raw C and CUDA, no PyTorch underneath
↳ github.com/karpathy/llm.c
6. nanochat -> a full chat model end to end
GPT-2 level training cost ~$43,000 in 2019
here it's about $48
↳ github.com/karpathy/nanochat
7. autoresearch -> agents that do the research
an agent edits the training code, trains for 5 minutes, keeps what improved and repeats all night
↳ github.com/karpathy/autorese…
train one GPT from scratch and you become the person who gets handed the AI project, along with the raise that comes with it
the classroom half of the same degree, what Stanford actually teaches about AI, is below ↓
Kirill 🔰 MBA retweeted
I validated it, NVIDIA did leave a free 60% on the table for Ampere/Ada GPUs 😎
Quad 3090s + Qwen 3.8 27B BF16 in vLLM:
Unpatched: 1114–1259 tok/s prefill
Patched: 1769–2020 tok/s prefill
FREE ~60%+ on prompt processing from 2K all the way to 128K. About 15 mins shaved off every hour for long running agentic tasks in Qwen 3.8 27B and other full precision dense models.
Decode, the same. Power utilization, the same. 🔥 risk, the same.
Quantized models + Imgen/Vidgen, testing soon to get hard numbers
Windows unlocked as well. 50X0, probably 🔜
Kernel module, no driver patching or vbios crap.
Easiest free win of the year for Local Ai 🎉
Notes + bench script: digitalspaceport.com/gpu-unl…
Kirill 🔰 MBA retweeted
github.com/impacifist/Window…
もうWindows版でてる開発が早すぎる
Kirill 🔰 MBA retweeted
And just like that,
We have windows gpu unlock
github.com/impacifist/Window…
Kirill 🔰 MBA retweeted
Meet the hero @CommonFoundry1
His work just unlocked 2x hidden Tensor Core performance on Ada GPUs (158 -> 309 TFLOPS on CUTLASS GEMM).
I dont have 40-series cards, but I tested the exact same boost on my RTX 3090s earlier this week. The hardware jump is completely real.
Follow him!
Common Foundry: We unlocked nearly 2× FP8→FP32 Tensor Core throughput on an RTX 4070 Ti SUPER. A matched GEMM in a custom NVIDIA CUTLASS 4.8.1 harness ran 1.95× faster: 158.54 → 309.08 TFLOP/s. Same GPU, binary and 240 W cap. Results:
github.com/JustAResearcher/g…
Kirill 🔰 MBA retweeted
Hands on AI Engineering!
I open-sourced a collection of 50+ hands-on AI engineering tutorials.
It features step-by-step projects and tutorials on:
• AI Agents and Multi-agents
• RAG (Agentic, Vision, and Local)
• MCP AI Agents
• OCR Apps
• Voice AI Agents
• & so much more
100% free and open source. 1k+ Github stars
I've shared the link in the comments!
Kirill 🔰 MBA retweeted
Happy 1024 & happy weekend, devs.
500M free tokens on ZCode: GLM-5.3 (200M) + GLM-5.3-Flash (300M). 200K packs, while they last.
Claim now. Use Oct 23 15:00 – Oct 26 01:00 UTC.
Update to the latest ZCode, claim in-app, then switch to Start Plan.
zcode.z.ai
Kirill 🔰 MBA retweeted
My RTX 3090 made a melancholic music video. Song, pictures, video, upscale. Not a single frame or note from a cloud model.
About 11 hours of pure generation time, including failed takes.
Video: Minimax H3, Minimax H3 Latent Upscaler
Images: Qwen2.1 Image, MOSS Image, Flux 3 Klein
Music: YuE2
Kirill 🔰 MBA retweeted
NVIDIA sold millions of 3090s with 50% of the tensor speed switched off. Not missing. Switched off.
1 kernel module flips it to full, and my BF16 tensor pipe went from 82 to 155 TFLOPS. Nothing else changed on the cards, and the outputs match the stock run byte for byte.
All so the gaming card wouldnt eat into workstation sales. Every 3090 owner ran at 50% for 6 years, so a pro card could look worth the price. Fine, its their business model.
And LLMs are the least interesting place for this. Decode waits on memory, and my engine runs its experts in int8. Still +10% on prefill.
The models that live on BF16 tensor math are everything else. Qwen Image 2.1 and FLUX 3 Klein, where every diffusion step is big matmuls. MiniMax H3, which takes 4 to 7 minutes on my cards. Music and speech models. And training, every LoRA anyone ever trained on a 3090 went through exactly the path that was cut.
Thats what I measure next. This is only the start.
My 3090s just got 1.9x FASTER at BF16 tensor math!!! Same cards, same driver, same 300W, one Linux kernel module.
bf16 tensor pipe: 82 -> 155 TFLOPS
cuBLAS BF16 GEMM 8192: 64 -> 99 TFLOPS
TF32: 41 -> 79 TFLOPS
NVIDIA ships every 3090 with fp32-accumulate tensor math at half speed. The driver even reports it: FMLA32 = 1/2. gpu-unlock by Zagitta sets it to full before the NVIDIA driver loads. After a cold boot all 4 of my cards say FMLA32 = full.
Real inference got less than I predicted. Strata, Qwen3.8 Flash-Next on 2 cards, prefill:
4K: +12%
10K: +12%
90K: 3205 -> 3536 tok/s, +10%
Before installing I said +14 to 19% at 90K, so I was wrong. The GEMM phases got 23% shorter, but they are only about 40% of Strata prefill, and the int8 experts, copies and CUDA core kernels didnt get any faster.
Outputs did not change. With the same expert cache size, my log-prob dump over 1996 teacher-forced positions is the same file byte for byte as the stock run before the module.
Next is Strata itself, now its everything around the GEMMs that makes prefill slow. How I set it up is in the replies.
Kirill 🔰 MBA retweeted
🇨🇳 One in five young Chinese can't find a job, and that's after removing the students from the math...
Kirill 🔰 MBA retweeted
Canadian rates are surging as the Bank of Canada warns about sharply higher inflation.
The problem is that Canadian households are drowning in debt.
Household debt stands at ~176% of disposable income, while housing prices have already fallen ~20% from their peak.
Higher rates mean higher mortgage payments, more pressure on housing prices and weaker consumption.
With debt levels this high, you can't fight inflation without crushing the economy.
