Bay Crest Consulting

tech consulting for companies worldwide.

Mumbai, India | @tarunn_b

[Delulu Check @delulucheck]

No more "asking for a friend" scenarios.

Stay tuned 👀

#DeluluCheck #RealityCheck #ComingSoon

[Teebs @Tarunn_b]

Elon Musk's side hustle of casually shooting rockets into space is worth $317 Billion, which is more than the entire net worth of the world's second-richest person (larry page at $265 bil).

[Teebs @Tarunn_b]

Vibe coding Fun Facts, Dec 2025

Credit: @arkhai_ai

[Teebs @Tarunn_b]

Quick glance on #Venezuela oil impact :

[Teebs @Tarunn_b]

PwC analyzed 1 billion job postings. Stanford tracked employment data since ChatGPT launched.

Here's what they found:

Workers WITH AI skills: +56% wage premium (doubled from 25% last year)

Entry-level workers in AI-exposed jobs: -20% employment since 2022

Same technology. Opposite outcomes.

AI job postings grew 7.5% this year.
Total job postings fell 11.3%.
Entry-level listings dropped 15%.

The pattern is clear:

AI doesn't replace workers.

It replaces workers who can't use AI.

The 22-year-old competing with ChatGPT for their first job? Losing.

The 35-year-old who learned to make ChatGPT do their grunt work? Getting a raise.

This isn't "AI taking jobs."

This is the biggest skills-based wage gap in a generation.

The divide isn't human vs machine.

It's humans who adapted vs humans who didn't.

And it's happening faster than anyone predicted.

Sources: PwC Global AI Jobs Barometer 2025 (n=1B job postings), Stanford Economics Working Paper Aug 2025, Handshake 2025

[Deedy @deedydas]

Venture Capital Compensation in the US

Many people kept telling me they don't know how VC comp works, so here it is, split by fund size, based on 2025 survey data, 500+ samples.

[Teebs @Tarunn_b]

I used 5.2 from release until now.
Generated long form detailed reports , chat , behavioral analysis , insightful/complex questions and i thought I'd end up writing a contrarian post compared to all the tech influencers once I saw the depth of it.

Nope, it's bullshit.

It's just a gpt 5.1 wrapper prompted to be a donkey.

[Teebs @Tarunn_b]

GPT 5.2 Released, here's everything you need to know:

- Fastest turnaround in OpenAI history:
GPT-5.2 dropped just 4 weeks after 5.1. Google's Gemini 3 topped benchmarks and Altman hit the panic button. This is the counter-punch.

- First AI to beat human experts at knowledge work:
GDPval benchmark covers 44 real occupations. Presentations, spreadsheets, schedules, diagrams. GPT-5.2 wins or ties against top industry professionals 70.9% of the time. GPT-5 was at 38.8%. Nearly doubled.

- 11x faster than humans at 1% of the cost:
Not hype. Actual measurements on professional knowledge tasks. This is the economic disruption everyone warned about.

- Competition math is solved:
AIME 2025 score: 100%. Perfect. No tools. GPT-5.1 was 94%. The problems that stump gifted high schoolers are now trivial.

- Abstract reasoning tripled:
ARC-AGI-2 is the hardest reasoning benchmark that exists. GPT-5.2 scores 52.9%. GPT-5.1 was 17.6%. Three times better in one release.

- Hallucinations cut by 30%:
Real ChatGPT queries tested. Response error rate dropped from 8.8% to 6.2%. This is what makes a model actually usable for work.

- Long context finally works:
4-needle test at 256k tokens. GPT-5.2 hits near 100% accuracy. GPT-5.1 collapsed to 40% at the same length. First model to ever do this.

- 8-needle test proves it:
Tracking 8 pieces of info across 256k tokens. GPT-5.2 holds at 75%. GPT-5.1 drops to 30%. The gap is brutal.

- Real document work is now possible:
Contracts, research papers, transcripts, multi-file codebases. Coherence across hundreds of thousands of tokens. This changes how professionals actually use AI.

- Investment banking spreadsheets:
Three-statement models, LBO models, proper formatting and citations. GPT-5.2 scores 68.4%. GPT-5.1 was 59.1%. Junior analysts should be nervous.

- Advanced math jumped hard:
FrontierMath Tier 1-3 tests expert-level problems. GPT-5.2 solves 40.3%. GPT-5.1 solved 31.0%. These problems took research mathematicians months to create.

- Science questions near-perfect:
GPQA Diamond is graduate-level physics, chemistry, biology. GPT-5.2 Thinking hits 92.4%. GPT-5.2 Pro hits 93.2%. No tools enabled.

- Scientific figure reasoning:
CharXiv benchmark for understanding charts from research papers. GPT-5.2 scores 88.7%. GPT-5.1 was 80.3%. Over 8 points better.

- Coding benchmark SWE-Bench Pro:
Tests 4 languages not just Python. Contamination-resistant. Real-world software engineering. GPT-5.2 hits 55.6%. GPT-5.1 was 50.8%.

- SWE-Bench Verified:
The established coding eval. GPT-5.2 hits 80%. GPT-5.1 was 76.3%. Steady march toward actually reliable code generation.

- Vision got a massive upgrade:
Error rates cut roughly in half on chart reasoning and UI understanding. Dashboards, screenshots, technical diagrams all dramatically better.

- Three model variants:
Instant for fast everyday stuff. Thinking for deep complex work. Pro for when quality matters more than speed.

- New xhigh reasoning setting:
Fifth level of reasoning effort for both Pro and Thinking. When you really need the model to cook.

- Context compaction for agents:
New /compact endpoint extends effective context window. Tool-heavy long-running workflows no longer hit the wall.

- Pricing:
$1.75 per million input tokens. $14 per million output. 90% discount on cached inputs. Higher per token but more efficient so actually cheaper per quality output.

- ChatGPT subscription unchanged:
Plus, Pro, Business, Enterprise users pay the same. No price increase.

- Enterprise partners already in production:
Notion, Shopify, Harvey, Zoom, Box all reporting state-of-the-art tool-calling and long-horizon reasoning.

- Coding tools onboard:
Cognition, Warp, JetBrains, Augment Code all say its their best model for agentic coding. Code reviews, bug finding, interactive coding all improved.

- Data science validated:
Databricks, Hex, Triple Whale found it exceptional for agentic data science and document analysis.

Continued in comments..

[Teebs @Tarunn_b]

GPT 5.2 Released, here's everything you need to know:

- Fastest turnaround in OpenAI history:
GPT-5.2 dropped just 4 weeks after 5.1. Google's Gemini 3 topped benchmarks and Altman hit the panic button. This is the counter-punch.

- First AI to beat human experts at knowledge work:
GDPval benchmark covers 44 real occupations. Presentations, spreadsheets, schedules, diagrams. GPT-5.2 wins or ties against top industry professionals 70.9% of the time. GPT-5 was at 38.8%. Nearly doubled.

- 11x faster than humans at 1% of the cost:
Not hype. Actual measurements on professional knowledge tasks. This is the economic disruption everyone warned about.

- Competition math is solved:
AIME 2025 score: 100%. Perfect. No tools. GPT-5.1 was 94%. The problems that stump gifted high schoolers are now trivial.

- Abstract reasoning tripled:
ARC-AGI-2 is the hardest reasoning benchmark that exists. GPT-5.2 scores 52.9%. GPT-5.1 was 17.6%. Three times better in one release.

- Hallucinations cut by 30%:
Real ChatGPT queries tested. Response error rate dropped from 8.8% to 6.2%. This is what makes a model actually usable for work.

- Long context finally works:
4-needle test at 256k tokens. GPT-5.2 hits near 100% accuracy. GPT-5.1 collapsed to 40% at the same length. First model to ever do this.

- 8-needle test proves it:
Tracking 8 pieces of info across 256k tokens. GPT-5.2 holds at 75%. GPT-5.1 drops to 30%. The gap is brutal.

- Real document work is now possible:
Contracts, research papers, transcripts, multi-file codebases. Coherence across hundreds of thousands of tokens. This changes how professionals actually use AI.

- Investment banking spreadsheets:
Three-statement models, LBO models, proper formatting and citations. GPT-5.2 scores 68.4%. GPT-5.1 was 59.1%. Junior analysts should be nervous.

- Advanced math jumped hard:
FrontierMath Tier 1-3 tests expert-level problems. GPT-5.2 solves 40.3%. GPT-5.1 solved 31.0%. These problems took research mathematicians months to create.

- Science questions near-perfect:
GPQA Diamond is graduate-level physics, chemistry, biology. GPT-5.2 Thinking hits 92.4%. GPT-5.2 Pro hits 93.2%. No tools enabled.

- Scientific figure reasoning:
CharXiv benchmark for understanding charts from research papers. GPT-5.2 scores 88.7%. GPT-5.1 was 80.3%. Over 8 points better.

- Coding benchmark SWE-Bench Pro:
Tests 4 languages not just Python. Contamination-resistant. Real-world software engineering. GPT-5.2 hits 55.6%. GPT-5.1 was 50.8%.

- SWE-Bench Verified:
The established coding eval. GPT-5.2 hits 80%. GPT-5.1 was 76.3%. Steady march toward actually reliable code generation.

- Vision got a massive upgrade:
Error rates cut roughly in half on chart reasoning and UI understanding. Dashboards, screenshots, technical diagrams all dramatically better.

- Three model variants:
Instant for fast everyday stuff. Thinking for deep complex work. Pro for when quality matters more than speed.

- New xhigh reasoning setting:
Fifth level of reasoning effort for both Pro and Thinking. When you really need the model to cook.

- Context compaction for agents:
New /compact endpoint extends effective context window. Tool-heavy long-running workflows no longer hit the wall.

- Pricing:
$1.75 per million input tokens. $14 per million output. 90% discount on cached inputs. Higher per token but more efficient so actually cheaper per quality output.

- ChatGPT subscription unchanged:
Plus, Pro, Business, Enterprise users pay the same. No price increase.

- Enterprise partners already in production:
Notion, Shopify, Harvey, Zoom, Box all reporting state-of-the-art tool-calling and long-horizon reasoning.

- Coding tools onboard:
Cognition, Warp, JetBrains, Augment Code all say its their best model for agentic coding. Code reviews, bug finding, interactive coding all improved.

- Data science validated:
Databricks, Hex, Triple Whale found it exceptional for agentic data science and document analysis.

Continued in comments..

[Teebs @Tarunn_b]

GPT 5.2 Released. Here's everything you need to know:

- Fastest turnaround in OpenAI history:
GPT-5.2 dropped just 4 weeks after 5.1. Google's Gemini 3 topped benchmarks and Altman hit the panic button. This is the counter-punch.

- First AI to beat human experts at knowledge work:
GDPval benchmark covers 44 real occupations. Presentations, spreadsheets, schedules, diagrams. GPT-5.2 wins or ties against top industry professionals 70.9% of the time. GPT-5 was at 38.8%. Nearly doubled.

- 11x faster than humans at 1% of the cost:
Not hype. Actual measurements on professional knowledge tasks. This is the economic disruption everyone warned about.

- Competition math is solved:
AIME 2025 score: 100%. Perfect. No tools. GPT-5.1 was 94%. The problems that stump gifted high schoolers are now trivial.

- Abstract reasoning tripled:
ARC-AGI-2 is the hardest reasoning benchmark that exists. GPT-5.2 scores 52.9%. GPT-5.1 was 17.6%. Three times better in one release.

- Hallucinations cut by 30%:
Real ChatGPT queries tested. Response error rate dropped from 8.8% to 6.2%. This is what makes a model actually usable for work.

- Long context finally works:
4-needle test at 256k tokens. GPT-5.2 hits near 100% accuracy. GPT-5.1 collapsed to 40% at the same length. First model to ever do this.

- 8-needle test proves it:
Tracking 8 pieces of info across 256k tokens. GPT-5.2 holds at 75%. GPT-5.1 drops to 30%. The gap is brutal.

- Real document work is now possible:
Contracts, research papers, transcripts, multi-file codebases. Coherence across hundreds of thousands of tokens. This changes how professionals actually use AI.

- Investment banking spreadsheets:
Three-statement models, LBO models, proper formatting and citations. GPT-5.2 scores 68.4%. GPT-5.1 was 59.1%. Junior analysts should be nervous.

- Advanced math jumped hard:
FrontierMath Tier 1-3 tests expert-level problems. GPT-5.2 solves 40.3%. GPT-5.1 solved 31.0%. These problems took research mathematicians months to create.

- Science questions near-perfect:
GPQA Diamond is graduate-level physics, chemistry, biology. GPT-5.2 Thinking hits 92.4%. GPT-5.2 Pro hits 93.2%. No tools enabled.

- Scientific figure reasoning:
CharXiv benchmark for understanding charts from research papers. GPT-5.2 scores 88.7%. GPT-5.1 was 80.3%. Over 8 points better.

- Coding benchmark SWE-Bench Pro:
Tests 4 languages not just Python. Contamination-resistant. Real-world software engineering. GPT-5.2 hits 55.6%. GPT-5.1 was 50.8%.

- SWE-Bench Verified:
The established coding eval. GPT-5.2 hits 80%. GPT-5.1 was 76.3%. Steady march toward actually reliable code generation.

- Vision got a massive upgrade:
Error rates cut roughly in half on chart reasoning and UI understanding. Dashboards, screenshots, technical diagrams all dramatically better.

- Three model variants:
Instant for fast everyday stuff. Thinking for deep complex work. Pro for when quality matters more than speed.

- New xhigh reasoning setting:
Fifth level of reasoning effort for both Pro and Thinking. When you really need the model to cook.

- Context compaction for agents:
New /compact endpoint extends effective context window. Tool-heavy long-running workflows no longer hit the wall.

- Pricing:
$1.75 per million input tokens. $14 per million output. 90% discount on cached inputs. Higher per token but more efficient so actually cheaper per quality output.

- ChatGPT subscription unchanged:
Plus, Pro, Business, Enterprise users pay the same. No price increase.

- Enterprise partners already in production:
Notion, Shopify, Harvey, Zoom, Box all reporting state-of-the-art tool-calling and long-horizon reasoning.

- Coding tools onboard:
Cognition, Warp, JetBrains, Augment Code all say its their best model for agentic coding. Code reviews, bug finding, interactive coding all improved.

- Data science validated:
Databricks, Hex, Triple Whale found it exceptional for agentic data science and document analysis.

Continued in comments..

[Teebs @Tarunn_b]

I asked claude to define what happiness is for humans.

Claude:

Happiness is not a destination or a feeling to chase. It's a byproduct of alignment between who you are and how you live.

My singular definition:

Happiness is the sustained sense that your days are spent in ways that matter to you, with people who see you, doing things that engage your full self.

--

Three pillars hold this up.

AGENCY — The felt sense that your choices shape your life. That you're the author not a passenger. The most miserable people are those who feel trapped.

CONNECTION — Not just any connection. The experience of being known and chosen anyway. You need at least one relationship where the full mess of you is visible and welcomed.

ENGAGEMENT — Something that absorbs you completely. Makes time disappear. The domain doesn't matter. The absorption does. The absence of this is what people mean when they say they're bored with life. It's not boredom. It's the horror of an unused self.

--

Here's the twist. The direct chase of happiness is the most reliable way to prevent it.

Instead it's constructed through smaller decisions.

--

Know what you actually value. Not what you've been told to value. Most people chase proxies. Money for security. Status for belonging. The proxies never satisfy because they're not the real thing.

Subtract before you add. Happiness is often blocked rather than absent. Removing one toxic relationship creates more happiness than adding five pleasant acquaintances.

Act before you feel ready. The feeling is downstream of the action. Not upstream. Waiting to feel happy before living differently is the trap.

--

Can anyone achieve it? Almost anyone. But prerequisites must be addressed first.

Basic physiological stability. Untreated mental illness and chronic pain consume all available attention. The mind cannot orient toward meaning when drowning in survival signals.

Minimum safety. You cannot plant seeds when you're running from a fire.

Capacity for honest self reflection. Some people are so defended against seeing themselves clearly they cannot make the adjustments happiness requires.

--

Given these prerequisites: yes. Available to the anxious, the traumatized, the difficult, the broken and rebuilt. But availability is not ease. Some have a longer road.

--

The paradox: happiness requires accepting unhappiness. The demand to feel happy all the time is itself a source of unhappiness. Some days are dark. Some losses cannot be fixed. Accepting this is not defeat. It's maturity.

--

Happiness is not sustained bliss. It's more like a resting heart rate. Not what you feel at any moment but the baseline you return to.

The happy life has frustration, sadness, anger, grief. But these states visit rather than reside.

--

Final truth.

Happiness is not something you find. It's something you build. And building it is the work of a lifetime.

The question is not how do I become happy. It's what can I do today that moves my life one degree closer to alignment.

Then tomorrow ask again.

That's everything.

[Teebs @Tarunn_b]

What are y'all using to code? 🤔

[Teebs @Tarunn_b]

Copy your entire git diff into Claude before asking for help.

Not the error. Not the file. The diff of everything that changed since it last worked.

"Here's every change made in the last 2 hours. Something in here broke it. Find it."

Claude can now binary search through your own mistakes instead of guessing at the codebase blind.

This is how you debug at 10x speed. You're not asking Claude to understand your whole system.

You're asking it to spot the one line that killed it in a constrained set of changes.

The bug is always in the diff. Always.

[Teebs @Tarunn_b]

Before you start any big fix, ask Claude to write a failing test for the bug first.

Not the fix. The test.

Now Claude has to prove it understands the problem deeply enough to define what "fixed" even means. If it can't write a test that fails correctly, it doesn't understand your bug.

And once that test exists, you run it after every change. Claude can't gaslight you that it's fixed when it's not.

You just made the AI accountable to something other than vibes.

[Teebs @Tarunn_b]

THE BULL CASE FOR SHORTING OPENAI

I spent 3 weeks investigating OpenAI's financials. What I found shocked me.

This isn't FUD. This is math.

Here's why OpenAI at $300B-$500B is the most dangerous mispricing in tech history. 👇

---

1. THE HEADLINE ILLUSION

OpenAI claims $13B in "annualized recurring revenue."

But here's what they don't tell you:

That's just one month × 12.

Leaked documents show Microsoft's 20% revenue share payments imply actual 2024 revenue was ~$2.5B—not the $4B reported.

The numbers don't reconcile.

---

2. THE BURN RATE FROM HELL

In 2024, OpenAI:
• Made ~$4B in revenue
• Lost $5B+ after costs
• Spent $2.25 for every $1 earned

That means they LOST money on every single paying customer.

More users = MORE losses.

This isn't scaling. This is bleeding.

---

3. THE $115 BILLION BURN PROJECTIONS

OpenAI's own documents project $115B in cumulative cash burn through 2029.

Their cash burn projections have already been REVISED UPWARD by $80B from earlier estimates.

In 2028 alone, they project $74B in operating losses.

Seventy-four. Billion. Dollars.

---

4. THE "PROFITABLE BY 2030" FANTASY

OpenAI told investors they'll hit $200B revenue and turn profitable by 2030.

HSBC just ran the numbers.

Their conclusion: Even at $200B revenue, OpenAI still needs $207B MORE in funding just to stay afloat.

They're committed to $1.4 TRILLION in compute spending through 2033.

The math doesn't work.

---

5. THE MICROSOFT TAX

Everyone forgets: Microsoft gets 20% of OpenAI's revenue.

In 2024 alone, OpenAI paid Microsoft ~$494M in revenue share.

Through Q3 2025: $866M.

OpenAI's "revenue" isn't really OpenAI's revenue.

And inference costs? Potentially HIGHER than their total revenue.

---

6. THE VALUATION ABSURDITY

$300B valuation on ~$4B 2024 revenue = 75x revenue multiple.

For context:
• Apple trades at ~8x
• Microsoft at ~12x
• Even peak-hype Tesla hit ~25x

OpenAI needs to 13x their revenue in 5 years just to reach their OWN projections.

That's 93% growth. Every. Single. Year.

---

7. THE NONPROFIT CONVERSION CHAOS

OpenAI was founded as a nonprofit "to benefit humanity."

Now they're desperately trying to convert to for-profit by Dec 2025.

Why? Because their $40B funding round REQUIRES it.

If conversion fails: the funding could convert to DEBT.

Elon Musk, Meta, and California's AG are all fighting to block it.

---

8. THE BRAIN DRAIN

Since 2024:
• Co-founder Ilya Sutskever (Chief Scientist) — GONE
• CTO Mira Murati — GONE
• Chief Research Officer Bob McGrew — GONE
��� Safety team co-lead Jan Leike — GONE

Their "Superalignment" safety team? Disbanded.

20+ key executives left in 18 months.

When the builders leave the building...

---

9. THE COMPETITION PROBLEM

2023: OpenAI had 50% enterprise market share.
2024: Dropped to 34%.
2025: Crashed to 25%.

Anthropic now LEADS enterprise AI at 32%.

DeepSeek offers comparable performance at 1/10th the cost.
Google Gemini just topped benchmarks.
Claude dominates coding with 42% share vs OpenAI's 21%.
Meta's Llama is free and open-source.

OpenAI's moat? Gone.

---

10. THE PONZI PATTERN

Look at the investor timeline:

• Oct 2024: $157B valuation
• March 2025: $300B valuation
• Oct 2025: $500B secondary valuation

Each round exists to mark up the previous round.

New money comes in. Old investors book paper gains. Rinse. Repeat.

SoftBank (lead investor) is financing their stake with LOANS.

---

11. THE "ANNUALIZED" SHELL GAME

OpenAI reports $12-13B "ARR" in 2025.

But The Information reports actual H1 2025 revenue was $4.3B.

For the math to work, OpenAI needs ~$2B/month by year-end.

Their growth rate is DECELERATING: 12.7% → 9.5% month-over-month.

They're not hitting their own targets.

-
Continued in comments

[Teebs @Tarunn_b]

Razor sharp focus. Zuckerberg understood value deeply.

[Teebs @Tarunn_b]

Peter Thiel and Elon Musk were driving to the meeting that would merge their companies into PayPal.

Musk was in his McLaren F1.
Million dollar car.

Thiel asked "what can this thing do?"

Musk floored it. Lost control. Car went spinning into an embankment on Sand Hill Road. Totaled.

They got out. Unhurt.

Thiel got a ride to Sequoia. Musk waited for the tow truck. Showed up later.

Never mentioned the crash.

Closed the deal anyway.

PayPal was born.

[Teebs @Tarunn_b]

In 2008, Airbnb was dying.

Brian Chesky had $30,000 in credit card debt. No investors would touch them. The idea of strangers sleeping in your house sounded insane.

So they pivoted. To cereal.

They designed custom boxes: "Obama O's: The Breakfast of Change" and "Cap'n McCains: A Maverick in Every Bite."

Bought the cheapest cereal they could find. Hot-glued 1,000 boxes shut by hand in their apartment.

Sold them for $40 each as "limited edition collector's items."

Made $30,000. Paid off their debt.

Paul Graham saw this and said: "If you can convince people to pay $40 for a $4 box of cereal, you can probably figure out how to make this company work."

He let them into Y Combinator.

Today Airbnb is worth $75 billion.

The cereal boxes are now in the Smithsonian.

[Teebs @Tarunn_b]

The 3-doc system that'll save you 20 hours of refactoring

Takes 30 minutes upfront. Saves your sanity.

Look, you can absolutely just start building. But you're gonna hit that moment at 2am where nothing connects and you're rewriting auth for the third time.

You're creating 3 documents before you write any code. Think of them as contracts between you and Claude.

DOCUMENT 1: DESIGN STYLE GUIDE

Do this first. Not kidding.

Everyone wants to start with features. But if you don't lock in the visual language early, you'll end up with Frankenstein UI. Every component looking like it's from a different app.

Open a chat. Tell Claude what vibe you want. Be specific. "I want it to feel like Linear meets Spotify. Dark mode. Minimal. Lots of breathing room. Subtle animations."

Then ask for:
- Color palette (background, surface, border, text primary, text secondary, accent, destructive)
- Font stack (what's the display font, what's the body font)
- Spacing scale (4, 8, 12, 16, 24, 32, 48, 64)
- Border radius tokens (none, sm, md, lg, full)
- Shadow tokens (if any)
- Animation timing (what's the default easing, default duration)

Get Claude to output this as a single doc. Save it. This is your bible.

Every time you build a component, you paste this in. No more "make it look nice." You say "follow the style guide."

DOCUMENT 2: FRONTEND PRD

Now you know what it looks like. Time to define what it does.

List every page. For each page, list:
- What the user sees (components, layout)
- What the user can do (actions, buttons, forms)
- What happens when they do it (navigation, state changes)

Don't write code. Write behavior.

Example:
"Dashboard page. Shows a list of projects as cards. Each card shows project name, last edited date, and a dropdown menu with edit/delete. Clicking a card opens the project. Delete asks for confirmation."

Then list your components. Just the names and what they do.
- ProjectCard: displays single project, handles click and dropdown
- ConfirmModal: reusable confirmation dialog
- EmptyState: shows when no projects exist

Finally: global state. What needs to be accessible everywhere?
- User auth state
- Current project
- Theme preference

Save this doc. When you're building frontend, this is what you reference.

DOCUMENT 3: BACKEND PRD

Last one. This is the engine.

Start with your data models. What are the things? What properties do they have?

User: id, email, name, avatar, created_at
Project: id, name, owner_id, created_at, updated_at

Then your API routes. Every endpoint. What it does. What it expects. What it returns.

POST /auth/signup - expects email, password. Returns user object and session.
GET /projects - expects auth header. Returns array of user's projects.
DELETE /projects/:id - expects auth header and project id. Returns success or error.

Then your business logic. The rules.
- Users can only see their own projects
- Deleting a project soft-deletes it for 30 days
- Email must be verified before creating projects

Save it.

HOW TO USE THESE

When you start building, you don't paste all three every time. You paste what's relevant.

Building a component? Paste the style guide.
Building a page? Paste the frontend PRD section for that page plus the style guide.
Building an API route? Paste the backend PRD.
Connecting frontend to backend? Paste both.

Claude now has exactly the context it needs. Not more, not less.

THE REAL UNLOCK

These docs aren't just for Claude. They're for you.

Writing them forces you to actually think through what you're building. You'll catch shit early. "Wait, how does a user actually get to this page?" "What happens if they're not logged in?"

That's the stuff that turns into 3am debugging sessions when you skip it.

30 minutes of docs. 20 hours saved. That's the trade.

Now go write them.

[Teebs @Tarunn_b]

How to not fuck up your database on day one:

The database is where most AI-built apps turn into spaghetti.

You're vibing, shipping features, and then suddenly you've got 14 tables that don't make sense and a users table with 47 columns.

Here's how to avoid that using Supabase MCP with Claude Code.

FIRST: WHAT EVEN IS MCP

MCP lets Claude actually see and interact with your Supabase database directly. Not you describing it. Not pasting schemas. Claude can look at your tables, run queries, check what exists.

It's like giving Claude the keys to the house instead of describing the furniture over the phone.

SET IT UP BEFORE YOU BUILD ANYTHING

Go to Supabase. Create your project. Get your project URL and service key.

Add the Supabase MCP to Claude Code. Now Claude can see your database in real time.

This changes everything. Because now instead of Claude guessing what your schema looks like, it can just check.

THE WORKFLOW THAT ACTUALLY WORKS

Before writing any application code, have a conversation about your data.

"I'm building an app where users can create projects and invite collaborators. What tables do I need? What are the relationships?"

Let Claude think through it. It'll propose a schema. Users, projects, project_members, maybe invites.

But here's the key: don't just let it create tables randomly as you build features.

Say: "Before we create anything, show me the full schema you're proposing. All tables, all columns, all relationships."

Review it. Poke holes. "What happens when a user is deleted? What about duplicate invites?"

Then: "Create these tables in Supabase."

Claude does it directly through MCP. You can verify in your Supabase dashboard.

Now you have a real schema before a single line of app code exists.

WHY THIS MATTERS

Most people do it backwards. They build a feature, realize they need a table, add it. Build another feature, add more columns. Repeat until the database is a crime scene.

With MCP, you can:

Check what exists before adding anything. "Show me my current tables" before every major feature. Claude sees reality, not its memory of what it built 47 messages ago.

Catch conflicts early. "I want to add a teams feature. How does this fit with the existing project_members table?" Claude can look at what's there and propose something that actually fits.

Run sanity checks. "Query the users table and show me what's in there." Make sure your migrations actually worked. Make sure data looks right.

THE ACTUAL COMMANDS

Once MCP is connected:

"Show me all tables in my database" - Claude lists everything
"Show me the schema for the projects table" - see columns, types, constraints
"Create a new table called invites with these columns..." - Claude creates it
"Add a foreign key from projects to users" - Claude handles the SQL
"Run this query and show me results" - test your data directly

You never have to guess what your database looks like. You never have to trust that a migration worked. You can just ask.

THE RULES I FOLLOW

Never create a table mid-feature. If I realize I need a new table, I stop. I think through the schema. I check how it relates to existing tables. Then I create it intentionally.

Check before you build. Starting a new feature? "Show me the current schema" first. Make sure you're working with reality.

Name things consistently from day one. Users not user. Project_members not projectMembers not members. Pick a convention and stick to it.

Add RLS policies early. "Add row level security to projects so users can only see their own." Don't leave this for later. You'll forget and ship an app where everyone sees everything.

THE UNLOCK

The database is the foundation. Everything sits on top of it. If it's messy, every feature you build inherits that mess.

Set it up. Plan your schema before you code. Check it often.

Your future self will thank you when you're not debugging foreign key nightmares at midnight.