Top Gun

[st1ne @SolSt1ne]

Michael Burry asked Washington one question in the New York Times - two weeks later the government came after him

he was a Stanford neurology resident posting stock picks at night - quit medicine in 2000 to run money

he read the actual mortgage documents nobody on Wall Street bothered to read - and saw the crash coming in 2005

his investors revolted - demanded their money back - he locked the fund and held the bet

the bet paid $700 million for his investors and $100 million for himself - his fund returned 489% while the S&P went nowhere

in 2010 he asked publicly: why didn't anyone in Washington listen to those who predicted the crisis?

"within two weeks, all six of my defunct funds were audited... and a little later, the FBI showed up"

then he stood in front of UCLA's economics graduates and told the whole story

21-min - free. watch it

bookmark and watch it today

[Ricardo @Ric_RTP]

The smartest man in AI just exposed the whole AGI narrative as a LIE.

And he used a physics problem from 1905 to prove it.

His name is Demis Hassabis. He runs Google DeepMind, and won the Nobel Prize for using AI to crack a problem in biology that had stumped scientists for 50 years.

Almost nobody in this industry has a track record like his.

He went on the NothingButTech podcast and called out the biggest lie in AI right now:

Right now the loudest voices in AI are telling you that AGI is basically here. OpenAI has literally defined AGI as a system that can outperform humans at most "economically valuable work." In other words, if it replaces enough jobs, we have arrived.

Hassabis thinks that bar is a joke.

He said real general intelligence has to do what the human brain can do, because the brain is the only proof we have that this kind of intelligence is even possible. He called that "a higher bar than just being able to do some useful economic work," which is about as close as a polite British Nobel laureate gets to calling his rivals out.

Then he gave the actual test:

Today's AI has read everything humans have ever written, including the theory of relativity. So when it explains relativity back to you, it's repeating an answer that already exists.

That's not intelligence.

So Hassabis proposed a test that makes memorization impossible. Train an AI on only what humanity knew in 1901, four years BEFORE Einstein published relativity. Then ask it to come up with relativity on its own.

It can't look up the answer, because in 1901 the answer doesn't exist yet. The only way to pass is to do what Einstein actually did: Take the same physics everyone else had and reason its way to an idea no human had ever had.

Hassabis says not a single AI today can, no matter how much it has memorized. Which means what we keep calling "almost AGI" is really just the best librarian in history.

It can find any answer that already exists but it cannot create one that doesn't.

His second version is even sharper:

AlphaGo, the system his own team built, famously invented a brand new move that no human had played in 2,000 years of the game.

Everyone called it genius but Hassabis says that still is not the bar.

The real test is not whether an AI can invent a new move inside Go, it is whether an AI could INVENT a game as deep and as beautiful as Go in the first place.

No model that exists today can do it.

The people telling you AGI has already arrived are the same people raising hundreds of billions of dollars on that exact promise.

The valuations only work if the finish line is right in front of us. So the finish line keeps getting dragged closer, and AGI keeps getting quietly redefined down to "does useful work," until the products they already sell happen to qualify.

Hassabis has nothing to prove and nothing to sell you. He already won the Nobel, and he is telling you the machines still cannot do the one thing that would make them genuinely intelligent, which is have a truly original idea.

To be fair to him, he is not a pessimist about it. He believes real AGI IS coming, and he is spending his life building it. He just refuses to pretend it is already sitting in your phone.

So the next time a founder tells you AGI is months away, remember that the one man in the room with a Nobel Prize built his test around Einstein, and admitted that nothing we have made can pass it.

What do you think?

[Top Gun @junktradeaim]

$MRLN

[BladudX @BladudX]

There are over 300,000 pilots in the world.

Only 8 are certified to land at Paro Airport in Bhutan. As the plane tilted 45° between the mountains, I finally understood why.🫪

[Cypherd @CypherdC]

$PRA California isn't known for being pro-business. Insurance regulators still reviewing this buyout where the combined co would control ~60% of market share in the state doesn't bode well for approval odds. Only 1.37% below take out price is WAY too tight IMO.

[IPCisco @ip_cisco]

Cisco SSH Configuration!!!| IPCisco
.
Please Like & Retweet..:)
.
#ccna #cisco #network

[Mohammed Tawakkal Ahmed @Tawakkalah13_10]

IPV4 vs IPV6

ipv4
address size is 32 bits or 2^32 id addresses naming standard numerical ID address four lots of 3 digits number separated by full stops like 199.99.3.3.3 it uses
NAT for muiltipe devices can share same public ip address configuration manual and DHCP configuration header size is (20–60 bytes) its simple and widely supported smaller address size easy to configure in small networks but also limited addresses
needs NAT because addresses are exhausted
more network congestion ex many cats live inside one apartment building so they share one public address to access network .

ipv6
address size is 128 bits id addresses naming is alphaumeric address 8 lots of four character hexodecimal number separted by colons
2600:1400:f:5a3l:3abd4 every device have a unique ip and also auto configuration on device using slaac
header size is Fixed (40 bytes)
most unlimited addresses also no need for NAT better routing efficiency and faster packet processing
but also more complex notation
example every cat gets its own mansion and its own unique address to access network

Advantages of IPv4

widely deployed across the Internet
And easier for humans to remember and type.

Advantages of IPv6

Better support for IoT (Internet of Things) devices.And also more efficient routing tables.
And automatic address configuration through SLAAC. Increase or Improved security features.

Can both work together ??

[philosophy @philosophy36l]

Appitude Cheat sheet

[Anshul Future Tech @Anshul7974]

Data storage evaluation......

[Ahmad @TheAhmadOsman]

INCREDIBLE

The MOST COMPLETE GUIDE for understanding benchmarks and evals, and why training on them is intentionally misleading is now available online to read for free

Covers the fundamentals

- What machine learning is actually trying to measure (generalization vs memorization)
- Data roles and why splits must stay sacred
- Leakage types and benchmark contamination
- Why LLMs make contamination uniquely hard (web-scale + synthetic + discussion + agents)
- The full contamination pipeline and semantic duplicates
- A practical taxonomy of "training on the test set"
- Why public benchmarks age, saturate, and stop working

Then the practical standards for clean measurement

- Proper evaluation design for classical ML and for LLMs
- Protocol freezing, exclusion lists, and honest reporting
- The rigorous before/during/after hygiene checklist
- The 2026 standard for serious LLM evaluation
- Benchmark lifecycle management and public goods thinking
- What is not a cardinal sin and what is INTENTIONALLY MISLEADING

You should read this, and if you cannot now then you most definitely wanna bookmark it for later

The benchmarks / evals / test sets are the rulers. Don't bend them.

[Tech Fusionist @techyoutbe]

Day 30/30 - complete. ☸️✅

30 days. 30 infographics.
One complete Kubernetes learning journey.

This is what we covered:
Day 1 → Kubernetes Fundamentals & Architecture
Day 5 → Pods, YAML, Labels & Selectors
Day 10 → Deployments, Services & Application Exposure
Day 15 → Storage, ConfigMaps & Secrets
Day 20 → Networking, Ingress & Cluster Operations
Day 25 → Security, RBAC, Monitoring & Troubleshooting
Day 29 → Interview Prep & Real-World Scenarios
Day 30 → Best Practices, Projects & Your Next Steps

Thank you to everyone who followed, saved, shared, and learned along the way 🙏

The complete 30-Day Kubernetes Learning Plan is now available free on Substack.

PDF version is available on Gumroad.

Links in reply 👇

Now tell me:
What should the next learning plan be?
Comments below 👇

[Govardhana Miriyala Kannaiah @govardhana_mk]

If you want to become good at DevOps and Cloud in 2026, then learn these tech differences first.

1) SSH vs SSL – Which Scales Better?
↳ https://t.co/jcuptmYNLU

2) Conventional Vs Kubernetes CI/CD Pipelines
↳ https://t.co/0fo0r61rJD

3) Dockerfile vs Docker Compose: What You Should Know
↳ https://t.co/dyl0ScknyU

4) Kubernetes Operator vs Helm - Which One to Choose ?
↳ https://t.co/eAsFl5DTAX

5) AWS Internet Gateway vs NAT Gateway – Which One to Choose?
↳ https://t.co/wC9x7qOxT3

6) Kubernetes Ingress Vs Gateway API
↳ https://t.co/6nd4Wr1rPc

7) GitOps Push vs Pull Architecture – Which One to Choose?
↳ https://t.co/4MMWiehsyG

8) CloudFront Signed URL vs S3 Pre Signed URL - When to Use What?
↳ https://t.co/nim2INsAnn

9) HPA vs VPA vs KEDA - Kubernetes Autoscaling
↳ https://t.co/TI70qA0QYM

49K+ read my DevOps and Cloud newsletter: https://t.co/WBucLdwdsb

What do we cover:
DevOps, Cloud, Kubernetes, IaC, GitOps, MLOps

🔁 Consider a repost if this is helpful.

[Dan Nanni @xmodulo]

Securing a Linux system takes more than one tool. It requires layered defense 😎👇

Find high-res pdf versions of all my cybersecurity related infographics from https://t.co/3t6LHw8TIY

#linux #cybersecurity #infosec #informationsecurity #software

[𝐇𝐚𝐦𝐳𝐚 | Network Engineer (Aspiring) @Hamzaonchain]

This is a raspberry pi.

A credit-card sized single-board computer that runs full Linux and behaves like a real system, just scaled down into something you can wire directly into hardware.

Under the hood, it’s a SoC (system-on-chip) setup: CPU + GPU + RAM (depending on model) on a single board, with GPIO pins exposed so you can interface with the physical world—sensors, relays, motors, cameras, LEDs, all of it.

That’s why it shows up everywhere in modern tech.

You’ll find it in IoT systems where it reads sensor data and pushes it to the cloud.

In home labs where it runs services like DNS, VPNs, or lightweight servers.

In cybersecurity setups where it’s used for packet capture, network monitoring, or portable test environments.

And in edge computing, where it processes data locally instead of relying on a remote server.

It’s not powerful in the “desktop PC” sense—but it’s extremely useful because it bridges software and hardware cleanly. You write code, deploy it, and it interacts with the physical environment almost immediately.

That’s the real value: fast prototyping of real-world systems without enterprise-level cost or infrastructure.

[Dhanian 🗯️ @e_opore]

Docker is still leading the DevOps highway

But if you could only keep ONE DevOps tool in your stack, what would it be?

🐳 Docker
☸️ Kubernetes
🏗️ Terraform
⚙️ Jenkins
📊 Prometheus

Drop your pick below 👇

[Avid @Av1dlive]

Anthropic's head of product for the Claude platform, on stage in Tokyo:

"We give our agents the ability to dream : they inspect their own previous trajectories and identify how to
self-improve."

memory + skills + dreaming + outcomes. the actual self-improving stack, demoed live this week.

[if i had only 2 days to set it up from zero:]

→ watch the managed agents section of the keynote (minute 21–30). free.
→ give your agent file-based memory: one folder, plain markdown, read it before every task
→ do one workflow by hand, then turn the transcript into a skill. that's the skills layer.
→ schedule a nightly dream pass: agent reads its own week, promotes what worked into permanent rules

that's the whole stack. the creator of claude code

already runs his setup this way

i rebuilt it piece by piece, loops and all.

save the full rebuild below ↓

[Sumanth @Sumanth_077]

Train your own LLM from scratch!

A step-by-step repo that walks you through building and training a transformer model from scratch using PyTorch. From downloading training data all the way to generating text.

The architecture is built from the ground up following the original "Attention is All You Need" paper. MLP, single head attention, multi-head attention, transformer blocks, and the full transformer model - all coded and explained with detailed diagrams at each step.

Training data comes from The Pile - a diverse 825GB open-source dataset covering books, articles, code, websites, and more. The repo includes scripts to download it, preprocess and tokenize it using tiktoken, store it in HDF5 format, and feed it into training batches.

You can train a 13M parameter model on a single Colab T4 GPU. At 13M parameters the model starts generating proper grammar and coherent short sentences. For billion-parameter training you need at least an A100 or RTX 4090. The repo includes a full GPU compatibility table so you know exactly what's possible on your hardware.

Includes a complete SFT and RLHF guide as a separate notebook for taking your trained model further.

Key capabilities:

• End-to-end pipeline: data download → preprocessing → training → text generation
• Full transformer implementation from scratch with PyTorch
• Trains models from 13M to 2B+ parameters on a single GPU
• Training data from The Pile (825GB, 22 diverse datasets)
• Tokenization via tiktoken (r50k_base)
• SFT and RLHF guide included

100% open source.

I've shared the link in the replies!