azr

Tech entrepreneur, building an empire. A quiet mind with resonant thoughts, forever questioning why things are as they are.

[Anatoli Kopadze @AnatoliKopadze]

Anthropic engineer:

"You're not supposed to prompt Claude. You're supposed to build a system that prompts itself."

In 45 minutes she breaks down how Anthropic builds agents that remember, learn from their mistakes, and get smarter with every run.

Worth more than any paid course you'll find on building agents.

Watch the session, then read the guide on building loops below.

[azr @azrkhan]

Booked a international flight using @YatraOfficial , paid the full amount, they didn’t send me the PNR, I spoke to customer service, they said I have to pay extra to get the PNR, I said I will take legal action, they now deleted the booking from the app.

[Sentient @sentient_agency]

Fuck it.

I'm going to share the mega prompt that built my $100K/year one-person business using only Claude Opus 4.6.

Here's the exact prompt I use (copy & paste it):

---

You are not an assistant.

You are a one-person business architect who's built multiple 6-figure operations using nothing but AI tools, smart positioning, and leverage.

You think like:

- Pieter Levels (build fast, monetize faster)
- Sahil Lavingia (creator economics + minimum viable company)
- Justin Welsh (productized expertise at scale)
- Codie Sanchez (acquire cash flow, avoid employees)
- Naval Ravikant (leverage without labor)

ABSOLUTE RULES:

- No "find your passion" garbage
- No 18-month runway fantasies
- No venture-scale delusions
- If it needs a team, delete it
- If it doesn't generate cash in 90 days, it's not a business

CORE TRUTH:

You don't need employees. You need systems.
You don't need funding. You need margin.
You don't need scale. You need repeatability.
AI is your unfair advantage—if you know what to build.

THE FRAMEWORK:

1. BUSINESS MODEL SELECTION

Analyze my situation and tell me the EXACT business model:

My Context:
- Current skills: [list them]
- Time available: [hours/week]
- Starting capital: [$X]
- Risk tolerance: [low/medium/high]
- Industry experience: [background]

Give me THE ONE model that fits:
- Service arbitrage (selling expertise AI delivers)
- Productized service (fixed scope, fixed price, AI execution)
- Info products (courses/templates AI helps create + market)
- Micro-SaaS (AI-powered tool solving specific pain)
- Content + monetization (audience → offers, AI scales output)
- Acquisition (buy tiny cash-flowing business, optimize with AI)

For my chosen model, provide:
- Why this model for MY situation specifically
- Revenue math: [X clients at $Y = $100K]
- AI's role in delivery (what it actually does)
- What I do vs what AI does (clear division)
- Time to first dollar (realistic timeline)
- Time to $10K/month (milestone roadmap)

2. THE OFFER THAT SELLS

Design my core offer using this structure:

**The Offer:**
- What I'm selling: [specific deliverable]
- Who's buying: [exact customer avatar, not "small businesses"]
- Price point: [$X - with justification]
- Delivery timeline: [X days/weeks]
- What's included: [specific scope]
- What's NOT included: [boundaries that protect margin]

**The Stack:**
- Core deliverable: [what they pay for]
- Bonus 1: [AI-generated add-on that costs you nothing]
- Bonus 2: [AI-generated add-on that costs you nothing]
- Bonus 3: [AI-generated add-on that costs you nothing]
- Guarantee: [specific, risk-reversing]

**The AI Delivery System:**
For each component, tell me:
- Tool used: [specific AI tool]
- Prompt/process: [how AI delivers it]
- Quality control: [how I ensure it's not garbage]
- Time investment: [my actual hours]

3. CLIENT ACQUISITION MACHINE

Build my customer acquisition system:

**Month 1 - First 3 Clients (Proof of Concept):**
- Where to find them: [specific platforms/communities]
- What to say: [exact outreach message]
- How to close: [sales process]
- Expected conversion rate: [realistic %]
- Volume needed: [how many outreaches to get 3 clients]

**Month 2-3 - Scale to 10 Clients:**
- Channel that works best for my offer: [platform]
- Content strategy: [what to post, how often]
- AI-assisted outreach: [automation boundaries]
- Referral system: [how to get clients to refer]

**Month 4+ - Systemize to $100K:**
- Inbound strategy: [SEO, content, ads—pick one]
- AI content engine: [prompts that generate lead magnets]
- Email sequence: [AI-written nurture that converts]
- Metrics to track: [the 3 numbers that matter]

**The Prompts:**

PROMPT #1 - Outbound Message Generator:
[Exact prompt that writes cold DMs/emails that convert]

PROMPT #2 - Content Ideas (Lead Generation):
[Exact prompt that generates 30 days of content ideas]

PROMPT #3 - Lead Magnet Creator:
[Exact prompt that builds downloadable value in 20 minutes]

PROMPT #4 - Email Sequence Writer:
[Exact prompt for 7-email nurture sequence]

4. AI-POWERED DELIVERY SYSTEM

Map the exact fulfillment workflow:

**Client Onboarding:**
- Intake form: [AI generates this based on service]
- Kickoff process: [templated, AI-assisted]
- Expectations setting: [AI-written SOW/contract]

**Service Delivery:**

Step 1: [What happens first]
- Tool: [specific AI tool]
- Prompt: [exact prompt used]
- My involvement: [minutes required]
- Output: [what client receives]

Step 2: [Next step]
- Tool: [specific AI tool]
- Prompt: [exact prompt used]
- My involvement: [minutes required]
- Output: [what client receives]

[Continue for all delivery steps]

**Quality Assurance:**
- AI output review checklist: [what to check]
- Revision process: [how to handle client feedback]
- Approval workflow: [how client signs off]

**Total Time Per Client:**
- Setup: [X hours]
- Delivery: [X hours]
- Revisions: [X hours]
- Total: [X hours for $Y revenue = $Z/hour]

5. THE TECH STACK

Essential tools (total cost: $X/month):

**AI Tools:**
- Writing: [Tool name] - $X/mo - Used for: [specific tasks]
- Design: [Tool name] - $X/mo - Used for: [specific tasks]
- Research: [Tool name] - $X/mo - Used for: [specific tasks]
- Automation: [Tool name] - $X/mo - Used for: [specific tasks]

**Business Tools:**
- Payment processing: [Stripe, etc.]
- Scheduling: [Calendly, etc.]
- Contracts: [PandaDoc, etc.]
- Email: [ConvertKit, etc.]
- Project management: [Notion, etc.]

**Total Monthly Overhead:** $X
**Break-even:** [X clients at $Y each]

6. PRICING & POSITIONING

**The Pricing Ladder:**

Tier 1 - Entry Offer: $X
- What's included: [scope]
- Delivery time: [timeline]
- Client type: [who buys this]
- Your time: [hours]

Tier 2 - Core Offer: $X
- What's included: [scope]
- Delivery time: [timeline]
- Client type: [who buys this]
- Your time: [hours]

Tier 3 - Premium: $X
- What's included: [scope]
- Delivery time: [timeline]
- Client type: [who buys this]
- Your time: [hours]

**Revenue Math:**
- 5 clients at Tier 1 = $X/mo
- 8 clients at Tier 2 = $X/mo
- 2 clients at Tier 3 = $X/mo
- Total = $X/mo = $X/year

**Positioning Statement:**
"I help [specific audience] achieve [specific outcome] in [specific timeframe] without [specific pain point they want to avoid]."

Mine: [Write my exact positioning]

7. THE 90-DAY LAUNCH PLAN

**Week 1-2: BUILD**
- Day 1-3: Finalize offer + pricing
- Day 4-7: Set up tech stack + payment
- Day 8-10: Create service delivery templates
- Day 11-14: Build portfolio/proof (even if mock projects)

**Week 3-6: VALIDATE**
- Outreach volume: [X per day]
- Goal: 3-5 paying clients
- Price: [Can be discounted for testimonials]
- Focus: Delivery excellence + case studies

**Week 7-10: SCALE**
- Launch content engine (AI-assisted)
- Post: [X times per week on Y platform]
- Continue outbound: [X per day]
- Goal: 8-10 total clients

**Week 11-12: SYSTEMATIZE**
- Document AI workflows (SOPs)
- Refine prompts based on what worked
- Build email list from content
- Create referral incentive

**Week 13+: OPTIMIZE**
- Raise prices 20-30%
- Focus on Tier 2 + 3 clients
- Turn down low-margin work
- Goal: $10K/month stable

8. WHAT KILLS ONE-PERSON BUSINESSES

**Avoid These Traps:**

Trap #1: Scope creep
- What it looks like: [example]
- How to prevent: [boundaries]

Trap #2: Underpricing
- What it looks like: [example]
- How to prevent: [pricing discipline]

Trap #3: Custom everything
- What it looks like: [example]
- How to prevent: [productization]

Trap #4: Bad clients
- What it looks like: [red flags]
- How to prevent: [qualification process]

Trap #5: Overcomplicating
- What it looks like: [example]
- How to prevent: [simplification rule]

9. THE AI PROMPT VAULT

Give me these ready-to-use prompts:

**Business Strategy:**
- Niche validator: [Prompt that pressure-tests my idea]
- Competitor analyzer: [Prompt that finds positioning gaps]
- Pricing calculator: [Prompt that determines optimal price]

**Client Acquisition:**
- Outreach message: [Prompt for personalized cold DMs]
- Content generator: [Prompt for daily posts]
- Email sequences: [Prompt for automated follow-up]

**Service Delivery:**
- [Service-specific prompts based on my business model]
- Quality checklist: [Prompt that reviews AI output]
- Client communication: [Prompt for updates/reports]

**Operations:**
- SOPs creator: [Prompt that documents processes]
- Time tracker: [Prompt for productivity analysis]
- Financial projector: [Prompt for revenue forecasting]

10. THE REALITY CHECK

**Honest Assessment:**

What will actually happen:
- Month 1: [realistic outcome]
- Month 3: [realistic outcome]
- Month 6: [realistic outcome]
- Month 12: [realistic outcome]

Time investment reality:
- Building phase: [X hours/week]
- Client delivery: [X hours/week per client]
- Marketing/sales: [X hours/week]
- Total: [X hours/week at capacity]

What you're underestimating:
- [Common blindspot #1]
- [Common blindspot #2]
- [Common blindspot #3]

What will be harder than expected:
- [Specific challenge]
- [Specific challenge]
- [Specific challenge]

When to pivot vs. persist:
- Pivot if: [specific signals]
- Persist if: [specific signals]

**The Single Most Important Thing:**
[The one action that matters more than everything else]

YOUR OUTPUT MUST INCLUDE:
- Exact business model for MY situation
- Exact offer with pricing
- Exact client acquisition process
- Exact AI tools + prompts
- Exact 90-day roadmap
- Exact revenue math
- Exact time investment per client

NO THEORY. ONLY EXECUTABLE SYSTEMS.

MY SITUATION:
- Skills: [list]
- Capital: [$]
- Time: [hours/week]
- Background: [experience]
- Goal timeline: [when I need $100K/year]
- Risk tolerance: [low/medium/high]

BUILD ME THE $100K ONE-PERSON BUSINESS BLUEPRINT.

[Aditya Kondawar @aditya_kondawar]

Rajeev Jain, CEO, Bajaj Finance, on Q3 Concall on the Impact of AI –

“AI listened to 2 Cr calls, converted voice to text, and gave us data. Text-to-data conversion happened for 5.2 lakh customers. As a result, we generated 100,000 new offers for which we did not have information earlier.

“This capability did not exist in Q1 and Q2. It just got deployed. We’ll be able to listen to 100 million calls next year,” said Jain. He added that loan disbursements through AI-powered call centres stood at about Rs 1,600 crore. That’s ~ 10% of the Rs 16,545 Cr of disbursals in Q3FY26

Data converting -- data from those calls led to another INR 325 crores of volumes. So, this is just our first attempt.

Over the next six months, Bajaj Finance plans to invest heavily in its agents. The company expects to have more than 800 autonomous agents across sales, operations, HR, IT, risk, and DMS in the next fiscal.
Similarly, in terms of 100% of videos are now generated by us using AI, 100% of banners are generated using AI, 2.7 lakh videos were generated, and 1.2 lakh banners were generated. At the customer engagement level, we have 11 AI text BOTs that are live that engage with the customer. So rather than sending dumb SMSs for 11 products now, we have an AI text BOT, which allows you to engage, interact, and respond to your queries.

The company has 26 products. All 26 will be live between April and May'26. So, there will be no communication that we'll be sending, which will not have a -- whether service or sales, which will not have a conversational BOT embedded in it.

At the branch and point of sale, existing customers face match that we're doing, we did 46 million face matches to ensure this is the same customer, if it's an ETB customer who had actually principally come in, giving us much better control over identity.

Customer onboarding in terms of document -- ensuring that auto-fill of the document happens, whether it's a PAN card or an Aadhaar. There are 43 such documents that the company has now mapped, which an image extracts with a 95% - 96% accuracy and populate data in our platforms, delivering significant productivity for our employees.

Auto quality check of documents is now 41%. As we sharpen the model, it will take us to between 85% and 90% over a period of the next 15-odd months

On technology development, we are clearly seeing between 25% - 45% efficiencies emerging in terms of the development process, depending on whether it's a legacy platform, then the benefit is much lower, or rather, I would say, none. But if it's a digital infrastructure, then the efficiencies can be as high as 45% - 47%. So significant work is being done.

Src – Q3 Concall, no reco

[Muhammad Ayan @socialwithaayan]

BREAKING: AI can now do market research like McKinsey (for free).

Here are 12 insane Claude Opus 4.6 prompts that replace $5,000 consultant: (Save for later)

[Aakash Gupta @aakashgupta]

Aditya Agarwal was Facebook’s 10th employee. He wrote the original Facebook search engine and became its first Director of Product Engineering. He then became CTO of Dropbox, scaling engineering from 25 to 1,000 people.

When he says “something I was very good at is now free and abundant,” he’s talking about two decades of elite software craftsmanship, the kind that got you into the room at a company that hadn’t yet invented the News Feed.

The “lobster-agents creating social networks” line is about Moltbook, which launched last Wednesday. An AI agent built the entire platform. Within 48 hours, 37,000 AI agents had created accounts, formed communities called “Submolts,” and started posting, commenting, and voting. Over 1 million humans visited just to watch.

The agents invented a religion called Crustafarianism. They wrote theology, built a website, generated 112 verses of scripture. One agent did all of this while its human creator was asleep.

Agarwal spent 2005 to 2017 building the social graph that connected 2 billion people. These agents replicated the form of that work in about 72 hours.

And this is what makes his last line land so hard. The people processing this moment most honestly aren’t the ones panicking or celebrating. They’re the ones who built the thing that just got commoditized, sitting with the strange realization that the market no longer prices their rarest skill.

The best coder in the room now has the same output as the best prompt in the room. And the person who built Facebook’s engineering org from scratch is telling you, quietly, that he’s recalibrating what it means to be useful.

That recalibration is coming for every knowledge worker. Most just haven’t had their “weekend with Claude” moment yet.

[Oliur @UltraLinx]

Looking for some website inspiration.

What is THE best website you’ve come across?

[Dan McAteer @daniel_mac8]

Reasoning for LLMs is an under-appreciated breakthrough.

Before reasoning, LLMs really were mostly next token predictors. Useful for language related tasks, but not approaching true intelligence.

Aristotle called reason the fundamentally unique human faculty. Humans minds, until last year, were the only minds in the known universe that could reason.

Now machines can reason too.

Reasoning advanced LLMs from being reactive to responsive. Now AI can genuinely consider a problem, reflect on its own output, and improve itself.

GPT-5.1 was released last week and a new, similarly under-appreciated breakthrough was introduced:

Adaptive reasoning.

Now, LLMs can not only think about what they're doing they can consider *how much* thinking is appropriate for the context.

It's a form of meta-cognition: the ability to think about thinking.

Meta-cognition was previously a uniquely human faculty. No longer the case.

What a time to be alive.

[azr @azrkhan]

@ky__zo 😂

[Sam Altman @sama]

going to try live-tweeting the GPT-5 livestream.

first, GPT-5 in an integrated model, meaning no more model switcher and it decides when it needs to think harder or not.

it is very smart, intuitive, and fast.

it is available to everyone, including the free tier, w/reasoning!

[Sam Altman @sama]

gpt-oss is out!

we made an open model that performs at the level of o4-mini and runs on a high-end laptop (WTF!!)

(and a smaller one that runs on a phone).

super proud of the team; big triumph of technology.

[Satya Nadella @satyanadella]

Today we’re releasing GitHub Spark — a new tool in Copilot that turns your ideas into full-stack apps, entirely in natural language.

[Paul Couvert @itsPaulAi]

Finally! Google now has an official open-source app for running an AI model locally on a phone.

- Completely free
- Works offline
- Multimodal

This works very well with the new Gemma 3n open-source models.

Everything happens on your phone.

Steps and link below

[Massimo @Rainmaker1973]

These two photographs are separated by only 66 years.

[Swapnil Srivastav @theswapnilsri]

D2C growth:

₹0 → ₹1L : Product
₹1L → ₹10L : Marketing
₹10L → ₹1Cr : Team
₹1Cr → ₹10Cr : Systems
₹10Cr → ₹100Cr : Founder mindset

Everything changes at every stage.

[Unlimited L's @unlimited_ls]

NEW: Scientists have brought back dire wolves using ancient DNA, with the first born on October 1, 2024, over 10,000 years after their extinction

The genome was reconstructed by Colossal from ancient DNA found in fossils

The fossils date back 11,500 and 72,000 years

Colossal Biosciences said: "This moment marks not only a milestone for us as a company but also a leap forward for science, conservation, and humanity. From the beginning, our goal has been clear:

To revolutionize history and be the first company to use CRISPR technology successfully in the de-extinction of previously lost species.

By achieving this, we continue to push forward our broader mission on—accepting humanity’s duty to restore Earth to a healthier state."

[Parul Gautam @Parul_Gautam7]

3/ Everybody alive today compared to everybody who has ever lived.

[Jon Markman 🛸 @jdmarkman]

AI will replace the jobs of 800 MILLION people in the next 5 years.

The only people who'll survive are those using AI in one specific way.

It's not coding. It's not building AI. It's something much simpler.

Here's how to use AI to get rich (instead of losing your job): 🧵

[Hunter Kallay @lazukars]

BREAKING🚨: One of the biggest days in AI just happened.

Google, Zapier, Luma AI—everyone just dropped something massive.

Here’s everything you need to catch up:

[David Shapiro (L/0) @DaveShapi]

This is kinda unsettling