Bruce Mareri

💙 🖤Will tweet as I wish and suffer the consequences…so it goes.  → 𝕏

[Bruce Mareri @brucemareri]

am bored and hungry

[Hayyan @hayyantechtalks]

🚨 Breaking News: Claude has a feature called "Human Writing Mode."

You can use it to completely eliminate robotic-style text, bypassing AI detectors like a professional editor.

Here are 6 access prompts: 👇

[Harris @HarrisDecodes]

Delete almost every prompt you've ever saved.

I did it.

For years, I collected prompts like they were assets.

→ Writing prompts.
→ Research prompts.
→ Marketing prompts.
→ Coding prompts.

Eventually I had hundreds & thousands of them.

And one day I realized something:

I was spending more time searching for prompts than actually using AI.

So I built a giant master prompt instead.

That worked...

Until every new chat forced me to explain myself all over again.

→ My style
→ My preferences
→ My workflow
→ My rules

The AI forgot everything.

I kept repeating myself.

Again. And again. And again.

That's when I stopped optimizing prompts.

And started optimizing memory.

Today I maintain just 3 markdown files.

And they've completely replaced my prompt library.

1️⃣

Don't write the file yourself.
Let Claude interview you.
Open a fresh Claude Cowork session and paste:

"You are building my about-me.md file.

Interview me with 20 questions, one at a time using AskUserQuestion.

Push back on vague answers.

Compile everything into a clean about-me.md under 2000 words."

Spend 20 minutes answering honestly.

The magic isn't the output.

It's the questions.

Claude will uncover patterns about how you think that you've never properly documented before.

By the end, you'll have a surprisingly accurate personal operating manual.

2️⃣

Keep it short.

My first version was over 20,000 words.

Terrible idea.

→ More tokens
→ More noise
→ Slower outputs

I cut it to under 2,000 words.

The quality improved immediately.

✔️ Smaller
✔️ Sharper
✔️ More useful

3️⃣

Create three files.

1. about-me.md

→ Who you are
→ How you think
→ Goals
→ Preferences

2. my-voice.md
→ Writing style
→ Favorite phrases
→ Phrases you avoid
→ Real writing samples
3. my-rules.md

→ Show a plan first
→ Ask before executing
→ Never delete without approval
→ Follow my workflow

Three files.

That's it.

4️⃣

Put them inside a single folder.

Claude Cowork
└── About Me

Then tell Claude:

"I'm Harris. Always read the files in my About Me folder before every task. Use my voice and follow my rules."

✔️ Done.

Now every task starts with context.

My prompts became dramatically shorter.

Sometimes one sentence. Sometimes one line.
But the outputs became:
• More accurate
• More consistent
• More useful
• More me

The biggest lesson?

People don't need more prompts.
They need a system that helps AI remember who they are.

Three markdown files solved a problem that thousands of saved prompts never could.

[Not Jerome Powell @alifarhat79]

Me after buying one share of SpaceX on Friday

[Suraj Sharma @suraj_sharma14]

If I had 6 months to become an Agentic AI Engineer.

I'd do this.

Stage 1: Python + Async Foundations
asyncio, FastAPI, event-driven architecture, error handling, API integration patterns.

Stage 2: LLM Fundamentals for Agents
Context management, model routing, token economics, latency tradeoffs, failure modes.

Stage 3: Tool Calling + Structured Outputs
Pydantic validation, function calling schemas, error recovery, dynamic tool discovery.

Stage 4: Memory + State Management
Short-term buffers, long-term vector recall, context compression, cross-session sync.

Stage 5: Single Agent Workflows
ReAct loops, plan-and-execute, self-reflection, iteration limits, graceful degradation.

Stage 6: Multi-Agent Orchestration
LangGraph/CrewAI, supervisor patterns, message passing, conflict resolution, handoffs.

Stage 7: Human-in-the-Loop Systems
Uncertainty detection, approval gates, audit trails, resume logic, intervention points.

Stage 8: Evaluation + Quality Assurance
Automated eval harnesses, LLM-as-a-judge, regression testing, hallucination metrics.

Stage 9: Observability + Tracing
Distributed tracing (LangSmith/Arize), cost dashboards, latency monitoring, alerting.

Stage 10: Security + Guardrails
Prompt injection defense, output filtering, PII redaction, sandboxed execution, compliance.

Stage 11: Production Deployment
vLLM/SGLang, Kubernetes scaling, CI/CD for agents, canary releases, rollback strategies.

Stage 12: Open Source + Portfolio
Ship autonomous agents publicly, write architecture docs, record demos, contribute to libs.

Most people stay stuck watching tutorials.

Builders get hired.

(Bookmark it)

[Movez @0xMovez]

Claude Code creator:

"100% of our pull requests at Anrtopic are run by Claude Code. 80–90% of code review too.

The feature I’m using the most today is /loops. I’m not prompting Claude anymore - I’m building loops"

in 1-hour interview, Boris reveals his setup, which helps him build the #1 coding tool of this year.

Worth more than a $500 vibe-coding course.

[Ren @renstocks_]

Satya Nadella just posted something that validates the entire AI buildout thesis from the very top of the stack.

The model is commoditizing. The durable value is the learning loop a company builds on top of the model.

He splits it into two assets:

Human capital -- the knowledge, judgment, relationships, and pattern recognition of your people.

Token capital -- the AI capability the firm builds and owns.

He says the real opportunity is building a learning loop where human capital and token capital compound together.

If the model layer is commoditizing then the durable returns are not in the model makers. They are in the infrastructure that powers every company building its own loop. Compute. Memory. Interconnect. Power.

The full stack underneath the application layer.

The model wars will have winners and losers. The infrastructure underneath gets bought either way.

Bullish the AI buildout.

Every layer. If you want to understand them in detail, check out my Substack.

https://t.co/Wna5UzCOVT

[Jafar Najafov @JafarNajafov]

A 22-year-old computer science student in Rochester, New York got so frustrated with the way JavaScript worked differently in every browser that he wrote a library to fix it.

He released it for free at an unconference in January 2006. That library became the foundation of the web for an entire generation.
It now runs on 77 percent of the top 10 million websites on Earth. He never started a company around it. He gave it away and went to work on education and Japanese art.

His name is John Resig. The library is called jQuery.

Here is the story.

John was born on May 8, 1984 in Boston, Massachusetts. He grew up fascinated by computers and programming. He attended the Rochester Institute of Technology and studied computer science with concentrations in economics and psychology. During college he lived in Computer Science House, an on-campus living and project community where students build things for fun. His housemates voted him Member of the Year during his sophomore year.

The problem that consumed him at RIT was cross-browser JavaScript. In 2005 the web was a mess. Internet Explorer, Firefox, Safari, and Opera all handled JavaScript differently. Writing code that worked the same way in every browser was painful, repetitive, and fragile. Developers spent more time fighting browser inconsistencies than building features.

John spent years at RIT building personal tools and libraries to make JavaScript easier to write. He wanted to share them with the world in a clear and concise way. In 2005, while still a student, he started building jQuery.

In January 2006 he released it at BarCamp NYC, a tech unconference whose core rule was no spectators, only participants. He was 22 years old. The library was small, elegant, and solved the exact problem every web developer was struggling with. You could traverse the DOM, handle events, animate elements, and make Ajax calls in a few lines of code that worked the same way everywhere.

The adoption was immediate. jQuery was the only open-source JavaScript library at the time that shipped with documentation. Most libraries expected developers to read the source code. John's first hire was not a developer. It was a community manager, someone to help answer questions from developers who were adopting the library. He understood from the beginning that adoption was a human problem, not a technical one.

Drupal selected jQuery as a core component. WordPress built on it. Microsoft adopted it. Apple used it. Google, IBM, Amazon, and AOL incorporated it into their websites. Django and Rails integrated it. Mozilla built with it.

By the early 2010s jQuery was everywhere. It became the most deployed JavaScript library in history by a margin so large that the second place contender was not even close. As of recent measurements, jQuery runs on approximately 77 percent of the top 10 million websites.

John worked at the Mozilla Corporation from 2007 to 2011 as a JavaScript evangelist and tool developer. He authored two books, Pro JavaScript Techniques and Secrets of the JavaScript Ninja. He created Processing.js, a port of the Processing language to JavaScript. He created Sizzle, a standalone CSS selector engine. He created QUnit, a testing framework.

In 2011 he joined Khan Academy as Chief Software Architect. He has been there ever since, building the technology behind one of the largest free education platforms on Earth.

Then he did something nobody expected from a JavaScript legend. He became a scholar of Japanese woodblock prints.

John is a Visiting Researcher at Ritsumeikan University in Kyoto, studying Ukiyo-e, the Japanese art of woodblock printing from the 17th to 19th centuries. He built ukiyo-e .org, a comprehensive database and image search engine that lets anyone search across hundreds of thousands of prints from museums and collections worldwide. He applied the same engineering instincts that made jQuery work, simplify, document, and make accessible, to a centuries-old art form.

He was inducted into RIT's Innovation Hall of Fame in 2010. He lives in the Hudson Valley of New York.

A frustrated college student wrote a JavaScript library in his dorm room and gave it away for free.

It became the invisible foundation of the modern web.

He went to Kyoto to study woodblock prints.

[Jaynit @jaynitx]

Elon Musk explains his 5-step algorithm for solving any problem:

"The most common mistake of smart engineers is to optimize a thing that should not exist."

"I have this very basic first principles algorithm that I run as a mantra."

Elon breaks it down:

Step 1: Question the requirements.

"Make the requirements less dumb. The requirements are always dumb to some degree, no matter how smart the person who gave you those requirements. You have to start there, because otherwise you could get the perfect answer to the wrong question."

Step 2: Try to delete it.

"Try to delete the part or the process step entirely. If you're not forced to put back at least 10% of what you delete, you're not deleting enough. Most people feel like they've succeeded if they haven't been forced to put things back in. But actually they haven't, they've been overly conservative and left things in that shouldn't be there."

Step 3: Optimize or simplify.

"The most common mistake of smart engineers is to optimize a thing that should not exist. So you don't optimize until after you've tried to delete."

Step 4: Speed it up.

"Any given thing can be done faster than you think. But you shouldn't speed things up until you've tried to delete it and optimize it otherwise, you're speeding up something that shouldn't exist."

Step 5: Automate.

"And then the fifth thing is to automate it."

Elon explains why the order matters:

"I've gone backwards so many times where I've automated something, sped it up, simplified it, and then deleted it. I got tired of doing that. So that's why I have this mantra."

[Millie Marconi @MillieMarconnni]

A Chinese mathematician spent 7 years making sandwiches at Subway after his PhD, and at 58 solved a 150-year-old math problem nobody thought was solvable.

His name is Yitang Zhang. The problem is called the Twin Prime Conjecture.

He was born in Shanghai in 1955 and knew he wanted to spend his life on mathematics by the time he was nine years old. That year he found his own proof of the Pythagorean theorem. Nobody taught it to him. He just worked it out.

Then the Cultural Revolution arrived and took everything.

The Chinese government closed the schools. Zhang's father had political troubles with the Communist Party, so Zhang was sent to the countryside with his mother to work in the fields. He spent 10 years as a farm laborer. No high school. No classroom. No teacher.

He read math books in the fields when he could find them.

When the revolution ended, Zhang was 23. He sat the university entrance exam and got into Peking University, one of the most competitive mathematics programs in China. He finished his bachelor's degree, then a master's. The president of Peking University personally recommended him for a full scholarship at Purdue University in the United States.

He arrived at Purdue in 1985. He earned his PhD in 1991.

Then the second wall hit.

His relationship with his doctoral advisor collapsed. The advisor did not write him letters of recommendation. Without those letters, the academic job market was closed. Zhang applied. Nothing came back. He spent the years after his PhD working as an accountant, doing delivery work, sleeping in his car during the stretches when nothing else was available.

A friend eventually opened a Subway sandwich restaurant in Kentucky and offered him a job. Zhang took it. He kept the books and made sandwiches. A man with a PhD in mathematics from Purdue, working a Subway counter because the academic world had no place for him.

He did this for seven years.

He was finally hired as a lecturer at the University of New Hampshire in 1999. Not a professor. A lecturer. The lowest rung of the academic ladder, with no research funding, no graduate students, and no institutional support. He taught calculus to undergraduates and worked on mathematics alone in whatever time was left.

Most people would have stopped believing by then.

Zhang did not stop.

The Twin Prime Conjecture is one of the oldest unsolved problems in number theory. Twin primes are pairs of prime numbers separated by exactly two: 5 and 7, 17 and 19, 41 and 43. The conjecture predicts that these pairs never stop appearing no matter how far you go along the number line. Mathematicians had believed this for over 150 years. Nobody had been able to prove it.

The deeper version of the problem asks something slightly different. Not whether twin primes are infinite, but whether there is any finite gap between prime numbers that appears infinitely often. This is called the bounded gap problem. The best mathematicians in analytic number theory had been attacking it for decades. A landmark 2005 paper by three researchers came agonizingly close and still could not close it.

Zhang worked on it alone. No collaborators. No funding. No department seminars where he could road-test his ideas. He once said he would go to a friend's house and think in the garden for hours.

In 2012, during a visit to a friend's home in Colorado, something unlocked.

He submitted his paper to the Annals of Mathematics in April 2013. The Annals is the most prestigious mathematics journal in the world. Papers sit in review for months, sometimes years. The editors read Zhang's submission and immediately knew something was different. They sent it to the leading experts in analytic number theory for review.

It was accepted in three weeks.

The paper proved that there are infinitely many pairs of prime numbers separated by a gap of less than 70 million. Not two. Not the twin prime gap specifically. But a finite gap. For the first time in history, someone had proved that prime numbers keep coming back together, that the universe of numbers never lets them drift apart forever.

Peter Sarnak, one of the most respected mathematicians at the Institute for Advanced Study, said: "He is not a fellow who had done much before. Nobody knew him. His result was spectacular."

Zhang was 58 years old.

Within a year he had the MacArthur Fellowship, the Cole Prize, the Rolf Schock Prize, and a full professorship at UC Santa Barbara. The man who spent seven years at Subway was now one of the most celebrated mathematicians alive.

He said in an interview: "I was not lucky. Maybe it is more important for a person to make himself known to the public. But that was not so easy for me."

He was not complaining. He was just being precise.

The mathematics establishment has a quiet belief that great work happens young. The Fields Medal cuts off at 40. Most mathematicians who change the field do it in their thirties. Zhang proved his most important theorem at 58, after a decade of farm labor, seven years of sandwiches, and a decade of teaching calculus to freshmen with no one watching.

He did not beat the deadline.

He proved there was no deadline to beat.

[Satya Nadella @satyanadella]

https://t.co/vLmiBKTtX3

[Tech with Mak @techNmak]

A dev got so frustrated watching his AI agent write 500 lines for a 5-line problem that he built a fix.

He called it Ponytail. Named after the guy every team has - long ponytail, oval glasses, been there longer than the version control. You show him fifty lines; he looks at them, says nothing, and replaces them with one.

Now your agent does the same. Before writing anything, it looks for a reason not to.

80-94% less code. 47-77% cheaper. 3-6x faster.

The best code is the code you never wrote.

GitHub Repo: https://t.co/WnFp9YNY53

[Rahul @sairahul1]

Sam Altman:

"We're going to see 10-person billion-dollar companies pretty soon."

"If I were 22 right now, I'd feel like the luckiest kid in history."

Most people will read this, feel inspired for 3 minutes, and go back to what they were doing.

The ones who act will build an app studio this weekend.

One tool. 10 minutes. $10K/month.

This is the how ↓

[Riley West @rileywestreel]

Nate Herk (creator of AI Automation Society):

"Claude Code isn't just for engineers. It becomes an exec assistant that runs your ops, tracks your team, and does your research — and it gets smarter every week."

in 27 minutes, Nate Herk turns Claude Code into a personal chief of staff: one CLAUDE.md as its brain, an onboarding interview, and a growing library of skills tailored to your processes.

setup is one folder and one file. what he stacks on top is the part that'll make you rethink your workflow…

Watch the talk, then read the article below.

That’s worth more than a $500 course on agent engineering.

[Charly Wargnier @DataChaz]

🚨 @Karpathy predicted the power of the "LLM Wiki." Google just formalized it.

Meet Open Knowledge Format (OKF): a vendor-neutral standard for giving foundation models the curated context they need.

I can genuinely see this replacing Notion, Obsidian, or traditional wikis for developer teams, and the reason comes down to bookkeeping.

Traditional wikis fail because humans inevitably abandon the tedious work of updating them.

As Andrej Karpathy pointed out recently, LLMs don't get bored.

They don't forget to update a cross-reference, and they can touch 15 files in a single pass.

OKF standardizes the interoperability layer so agents can actually do that heavy lifting autonomously.

Because the format is minimally opinionated, it doesn't dictate what you write, it just dictates how it's structured. You get:
→ Human-readable documents that live right alongside your code in version control
→ Cross-links that map out complex entity relationships without needing a graph database
→ A system that survives moving between different tools and organizations

There is no complex compression scheme.

No central registry.

If you can cat a file, you can read it.

If you can git clone a repo, you can deploy it.

This is how we stop rebuilding context pipelines from scratch every time a new model drops.

Announcement + spec file in 🧵↓

[Stellarix @Stellarixorine]

Mathematics is a powerful weapon that change the world.

[Rimsha Bhardwaj @heyrimsha]

In 1986 an American programmer published the first article ever written about Windows programming. Two years later he published the book that taught the world how to write software for the operating system that would go on to run 90 percent of the computers on Earth. The phrase "look it up in Petzold" became shorthand among developers for three decades.

He is 73 years old, lives in a small town in upstate New York, and published a second edition of his most famous work two years ago.

His name is Charles Petzold.

Here is the story, because the person who taught two generations how to build software for the most widely used operating system in history has almost no public profile in the modern tech world.

Charles was born in 1953 in New Brunswick, New Jersey. He earned a Master of Science in Mathematics from Stevens Institute of Technology in 1975. In the early 1970s, before personal computers existed as a consumer product, he built a computer from scratch using a Z-80 processor and wired it to control a music synthesizer. He was programming machines before most people had ever seen one.

In the mid-1980s he began writing programs and technical articles for Microsoft Systems Journal. In 1986 he published what is recognized as the first article on Windows programming ever written. He served as a contributing editor at the journal from 1985 to 2000.

In 1988 Microsoft Press published Programming Windows. The book did something nobody else had done. It took the Windows API, a sprawling, undocumented, and deeply intimidating interface for building graphical software, and explained it from first principles with working C code. Every example compiled. Every explanation built on the last. The book assumed you knew C and nothing else.

It became the bible for Windows development. Six editions followed, tracking the evolution of Windows from 3.0 to 3.1 to 95 to 98 to XP to 8. The fifth edition alone ran to 1,479 pages. Generations of professional programmers learned their craft from this single book. "Look it up in Petzold" became the standard response in developer forums, in offices, and in computer science departments when anyone asked how to do something in Windows.

Microsoft named him a Windows Pioneer, one of only seven people to ever receive the award. He won the Microsoft Most Valuable Professional designation. He became the reference that the reference books referenced.

Then in 1999 he published Code: The Hidden Language of Computer Hardware and Software.

Code is a different kind of book. It starts with flashlights and Morse code and Braille. It builds, layer by layer, from relay switches to logic gates to flip-flops to registers to a working CPU. By the end of the book you understand how a computer works at the hardware level without needing an engineering degree. The writing is patient, clear, and free of jargon.

The idea for Code came to him in 1987 while writing a column called "PC Tutor." It took him twelve years to write. It became one of the most beloved technical books ever published. Scott Hanselman, a Partner Program Director at Microsoft, called it "the first book about programming that spoke to me" and said it taught him "how many unseen layers there are between the computer systems that we as users look at every day and the magical silicon rocks that we infused with lightning and taught to think."

In August 2022 Charles published the second edition of Code from his home in Roscoe, New York. He expanded the book, added new chapters, and built an interactive companion website at https://t.co/LGicky9iiv with animated circuit diagrams he made himself. He was 69 years old.

Between Programming Windows and Code he published over a dozen other books. Programming Windows Phone 7. The Annotated Turing, a guided tour through Alan Turing's 1936 paper on computability. Creating Mobile Apps with Xamarin.Forms. 3D Programming for Windows. Programming in the Key of C#. He worked at Xamarin from 2014 to 2018.

His personal website at charlespetzold .com is a flat page. His blog posts are written in plain HTML. There is no marketing copy. There is no newsletter. There is no podcast.

A mathematician from New Jersey taught the world how to program the most important operating system in history, then wrote the book that explained how computers actually work.

He is still writing from a small town in upstate New York.

[Akshay 🚀 @akshay_pachaar]

The top Claude Code CLI integrations to give you superpowers:

1. GitHub

The repo stops being a folder of files and becomes something the agent actually runs.

It reads and writes issues, PRs, Actions, and releases, so it works the codebase the way an engineer does, not by editing text on disk.

This is the gap between an agent that touches code and one that actually ships it.

2. HuggingFace

This is where your models and datasets live, and the agent can reach all of it.

It pulls a base model, runs the training, and pushes the fine-tuned version back, without you ever leaving the terminal.

The whole loop happens in one place.

3. Bright Data

Web access that actually works for an agent, instead of a scraper you have to babysit.

It pulls live search, full pages, and clean data from sites that normally block bots, and now it can even build custom scrapers from the terminal.

Collect data from any website by turning prompts into ready-to-run scrapers with built-in proxies and automatic unblocking.

GitHub: https://t.co/8OciAbQykA

(don't forget to star 🌟)

4. Stripe

Payments without ever opening the dashboard.

It forwards live webhooks and fires real payment events, so the agent runs through the whole checkout instead of faking it.

The only real way to know your billing works is to run actual money through it.

5. InsForge

A full backend in one CLI.

Database, auth, storage, edge functions, hosting, and an AI gateway, all in one place instead of stitching five services together. The agent sets up the infrastructure the way a backend engineer would.

Think of it as a backend built for agents.

GitHub: https://t.co/4pPPor1tyb

(don't forget to star 🌟)

6. CodeRabbit

It reviews the agent's own code before you ever see it.

It catches bugs, security holes, and sloppy patterns while the change is still local, so nothing messy makes it into a PR.

An agent that checks its own work first is a very different teammate.

7. Playwright

It gives the agent hands in a real browser.

Click, fill forms, take screenshots, and run UI tests across Chrome, Firefox, and Safari, on the real page instead of guessing from the HTML.

Easily the most underrated way to let an agent check what it built.

8. Google Workspace

Gmail, Drive, Calendar, Sheets, and Docs through one connector.

It is built on Google's own APIs and made for agents to actually do the work, not just read it, so it can draft the reply, update the sheet, and block off the calendar in one go.

An agent that can read your inbox but can't act on it is only half useful.

9. Slack

It puts the agent right where your team already works.

It builds and runs workflows that post updates and sort through channels, so "tell me what I missed in # incidents and flag anything urgent" just happens without you switching tabs.

This is the one that makes the agent feel present instead of stuck in a terminal.

10. E2B

A safe sandbox for code the agent wrote itself.

It spins up a small isolated VM, runs the code, grabs the output, then shuts the whole thing down.

This is what lets you actually let the agent run what it writes.

GitHub: https://t.co/rzvXVzzllo

(don't forget to star 🌟)

11. Unsloth

Fast local fine-tuning without the cloud bill.

It trains LoRA and QLoRA adapters about 2x faster on a lot less VRAM, then exports to GGUF or pushes straight to the hub.

This is what turns fine-tuning from a whole project into just another step.

GitHub: https://t.co/jUhjsrUZGD

(don't forget to star 🌟)

12. ffmpeg

The media tool that does almost everything, now in the agent's hands.

Cut, convert, and pull audio or video out of just about anything in a single command.

Old, unglamorous, and still the thing every media pipeline quietly runs on.

----

That said, if you want to see how this whole stack fits together, I wrote a full deep dive on how Claude Code's harness works, what actually goes in the .claude/ folder, and how hooks, skills, and subagents come together into a real workflow.

The article is quoted below.

[Riley West @rileywestreel]

https://t.co/6F8sV4ZHSM

[Alok Kumar @Alokkumarzz]

INSTEAD OF WATCHING AN HOUR OF NETFLIX TONIGHT.

This 60-minute Cambridge lecture by Demis Hassabis will teach you more about the future of AI than most people will learn in the next 5 years.

Bookmark it and give it an hour, no matter what.

[rari @0xwhrrari]

Anthropic and OpenAI are both telling engineers to write loops

Not prompts
Not agents
Loops

That is not a coincidence

When the two biggest AI labs converge on the same pattern, it is a signal

Most engineers still think in single calls

Input → Model → Output

The ones winning in 2026 think in cycles

Output becomes input
The model checks its own work
The loop runs until the result is right

Bookmark this before loops become the default workflow

Then read the article below