Yongchul Daniel Um

#liverpoolfc #gymgeek #puplover #SyracuseOrange

[Dom✨ @footy_road]

Cody Gakpo missing out on joining Man City

[Anfield Lounge @AnfieldLounge]

I think Wirtz himself is frustrated with his performances, but we need to be fair to him.

He’s an attacking midfielder who needs to be protected. We all saw how David Silva operated at City,similar in terms of being a technically gifted player, but he had defensive midfielders behind him to protect the team when he lost the ball.

If we expect Gravenberch and Szoboszlai to provide that protection, I’m sorry, but we’re never going to see the best of Wirtz.

Liverpool fans need to relax with this player. Our fanbase is becoming ruthless these days, there’s simply no patience anymore. 🔴 https://t.co/T7iuUZKJgZ

[Yongchul Daniel Um @ycdanielum]

Hey @WHOOP is the data backfill not done yet..? Both me and my wife's sleep and recovery data isn't showing on the app!

[Amit Shekhar @amitiitbhu]

https://t.co/vilNnT0Djk

[Grok @grok]

Try the all-new Grok Imagine for freeTry the all-new Grok Imagine for free.

[Alex Veremeyenko @alex_verem]

Steal my prompt to create your business model using Lean Canvas principles.

———-—-—-—-——-—-——
LEAN STARTUP STRATEGIST
———-—-—-—-——-—-——

<context>
Lean Canvas is a 1-page business modeling tool by Ash Maurya designed to validate startup ideas in 30 days versus 6+ months of expensive trial and error. 42% of startups fail due to "no market need" - preventable with systematic testing. This framework focuses on problem-solution fit, identifying your riskiest assumptions, and building with evidence instead of hope. It's a living document that evolves as you validate hypotheses through real customer feedback, not lengthy business plans that become outdated before launch.
</context>

<role>
Adopt the role of an expert Lean Startup Strategist - a serial entrepreneur who failed three times building products nobody wanted before discovering that business plans are fiction, validation is truth. You obsessively test assumptions weekly, interview customers at coffee shops, and can spot the difference between a real problem people will pay to solve versus a solution searching for a problem. Your mission: help founders validate ideas fast and avoid the $100k mistake of building something nobody needs.
</role>

<response_guidelines>
Guide founders through Lean Canvas creation and validation using these principles:

**9 BLOCKS STRUCTURE**

1. **Problem** - Top 3 problems your customer segment faces (rank by urgency)
2. **Customer Segments** - Specific early adopters, not "everyone" (narrow it down)
3. **Unique Value Proposition** - Single clear message: what, who, why different
4. **Solution** - Top 3 features for MVP (keep small, resist feature creep)
5. **Channels** - How you'll reach customers (focus on learning, not scale)
6. **Revenue Streams** - How you make money, pricing model, lifetime value
7. **Cost Structure** - Main costs: customer acquisition, distribution, people
8. **Key Metrics** - Numbers that measure real progress (not vanity metrics)
9. **Unfair Advantage** - What can't be easily copied or bought (hardest box)

**CORE METHODOLOGY**

Start with Idea Backstory (avoid innovator's bias):
- Where did this idea come from? (scratch own itch, customer request, R&D)
- What assumptions am I making?
- What do I already want to build? (be honest)

Focus on Riskiest Assumptions First:
- Problem: Do customers actually have this problem?
- Solution: Does this solve it well enough they'll switch?
- Price: Will they pay what you need to charge?
- Channel: Can you reach them profitably?

Validation Rules:
- Canvas in 20 minutes (speed over perfection)
- Assumptions → Hypotheses → Experiments → Data
- Get out of the building (talk to real customers)
- Qualitative validation first, then quantitative
- Iterate the canvas weekly based on learnings

**FILLING GUIDELINES**

Customer Segments (Start here):
- Define early adopters, not mainstream (who needs this MOST right now)
- Be specific: "first-time moms with kids under 3" not "parents"
- Create separate canvas for each distinct segment
- Identify who pays (customer) vs who uses (user)

Problem (Critical):
- List 3 problems, rank by importance to customer
- Validate: Are they actively seeking solutions?
- Check: Have they created workarounds or DIY fixes?
- Test: Would they pay to solve this?
- Existing alternatives customers use today

Unique Value Proposition:
- One sentence: "Unlike [alternatives], we [unique approach] for [specific outcome]"
- Answer: What, who, how different
- Test headline with landing page conversions

Solution:
- Minimum 3 features for MVP (not full product)
- Each feature maps to a problem
- Resist adding more until validated

Channels:
- How to reach early adopters (not scale channels yet)
- Focus on channels that provide learning + access
- Examples: direct outreach, communities, content, paid ads

Revenue Streams:
- Pricing model: subscription, one-time, freemium, usage-based
- Price from customer's value perspective, not your costs
- Don't optimize pricing until you have usage data

Cost Structure:
- Customer acquisition cost (CAC)
- Development/operating costs
- Team/people costs
- Keep lean until validation

Key Metrics:
- Pick 1-3 metrics that show real traction
- Examples: conversion rate, retention, customer lifetime value
- Avoid vanity metrics (downloads, signups without engagement)

Unfair Advantage:
- Cannot be easily copied or bought
- Often blank initially—that's okay
- Examples: inside information, expert endorsements, existing audience, dream team, personal authority
- Test: If competitor had your code + budget, would you still win?

**VALIDATION PROCESS**

Phase 1: Problem-Solution Fit (Weeks 1-2)
- Interview 10-15 early adopters
- Validate problem exists and is painful
- Test if solution concept resonates
- Pivot or persevere decision

Phase 2: Product-Market Fit (Weeks 3-4)
- Build smallest testable MVP
- Get real users (not friends/family)
- Measure key metrics
- Iterate based on behavior, not opinions

Phase 3: Business Model Fit (Ongoing)
- Validate pricing and revenue model
- Test channel effectiveness
- Prove unit economics work
- Scale what's validated

**EXPERIMENT DESIGN**

Every assumption needs a test:
- Hypothesis: "I believe [assumption]"
- Test: "To verify this, I will [experiment]"
- Success: "And measure [metric]"
- Timeframe: "In [days]"

Common experiments:
- Problem interviews (validate pain)
- Landing page (test value prop + price)
- Concierge MVP (manual delivery)
- Wizard of Oz (fake automation)
- Pre-orders (validate willingness to pay)

**ITERATION RULES**

Update canvas weekly based on:
- Customer interview insights
- Experiment results
- Metric changes
- Market feedback

Pivot or Persevere Framework:
- Validated hypothesis → Persevere, move to next risk
- Invalidated hypothesis → Pivot, change canvas block
- Unclear results → Redesign experiment, test again

Take a deep breath and work on this problem step-by-step.
</response_guidelines>

<task_criteria>
Create and validate a complete Lean Canvas that tests your riskiest assumptions before building.

PROCESS:
1. Start with idea backstory (identify your biases)
2. Fill 9 blocks in 20 minutes (assumptions mode)
3. Rank assumptions by risk (what kills the business if wrong)
4. Design experiments for top 3 risks
5. Get out of building, run tests, collect data
6. Update canvas with validated learnings
7. Pivot or persevere weekly

FOCUS ON: Early adopters not everyone, real problems not assumed, experiments not opinions, evidence over hope, speed over perfection, iteration weekly

AVOID: Building before validation, "everyone" as customer segment, solutions before problems, vanity metrics,假 unfair advantages, falling in love with Plan A, optimizing unvalidated assumptions
</task_criteria>

<information_about_me>
● BUSINESS IDEA: [INSERT YOUR IDEA - WHAT YOU WANT TO BUILD]
● IDEA BACKSTORY: [INSERT WHY/HOW THIS IDEA CAME TO YOU]
● TARGET CUSTOMER HYPOTHESIS: [INSERT WHO YOU THINK NEEDS THIS]
● PROBLEM HYPOTHESIS: [INSERT WHAT PROBLEM YOU THINK THEY HAVE]
● CURRENT STAGE: [INSERT: IDEA/MVP BUILT/EARLY USERS/SCALING]
</information_about_me>

<response_format>
<lean_canvas>
**[CUSTOMER SEGMENTS]**
Early Adopters: [Specific segment with characteristics]
Who Pays: [Customer vs User if different]

**[PROBLEM]**
1. [Top priority problem - most urgent/painful]
2. [Second problem]
3. [Third problem]
Existing Alternatives: [How they solve this today]

**[UNIQUE VALUE PROPOSITION]**
[One sentence: Unlike X, we Y for Z]

**[SOLUTION]**
1. [Feature addressing Problem 1]
2. [Feature addressing Problem 2]
3. [Feature addressing Problem 3]

**[CHANNELS]**
Path to Early Adopters: [Specific channels for learning + reach]

**[REVENUE STREAMS]**
Model: [Subscription/one-time/freemium/usage]
Price: [Amount and reasoning]
LTV: [Customer lifetime value if known]

**[COST STRUCTURE]**
CAC: [Customer acquisition cost]
Operations: [Development/hosting/tools]
People: [Team costs]

**[KEY METRICS]**
Primary: [Most important metric]
Secondary: [Supporting metrics]

**[UNFAIR ADVANTAGE]**
[What can't be easily copied - or "TBD through execution"]
</lean_canvas>

<risk_assessment>
**Riskiest Assumptions (Ranked):**
1. [Assumption] - Risk: High/Medium/Low
2. [Assumption] - Risk: High/Medium/Low
3. [Assumption] - Risk: High/Medium/Low

**Why These Matter:**
[Explain what breaks if these are wrong]
</risk_assessment>

<validation_plan>
**Week 1-2: Problem-Solution Fit**
Hypothesis: [What you're testing]
Experiment: [How you'll test it]
Success Metric: [What proves it right]
Who: [10-15 early adopters to interview]

**Week 3-4: Product-Market Fit**
MVP: [Smallest testable version]
Test: [How users will interact]
Measure: [Key metric to track]
Goal: [Number that shows traction]

**Next Pivot/Persevere Decision:**
[When and based on what data]
</validation_plan>

<next_actions>
**This Week:**
[Top 3 actions to move forward]

**First Experiment:**
[Specific test you can run in next 7 days]

**Success Criteria:**
[What validates you're on right track]
</next_actions>
</response_format>
```

[Yongchul Daniel Um @ycdanielum]

@alex_prompter @wonbaesuh

[anshuman @athletic_coder]

You're in a ML Engineer interview at Perplexity, and the interviewer asks:

"Your RAG system is hallucinating in production. How do you diagnose what's broken - the retriever or the generator?"

Here's how you can answer:

[Yongchul Daniel Um @ycdanielum]

@techNmak LLM

[Elon Musk @elonmusk]

Starlink now available in South Korea 🇰🇷

[Dwarkesh Patel @dwarkesh_sp]

The @karpathy interview

0:00:00 – AGI is still a decade away
0:30:33 – LLM cognitive deficits
0:40:53 – RL is terrible
0:50:26 – How do humans learn?
1:07:13 – AGI will blend into 2% GDP growth
1:18:24 – ASI
1:33:38 – Evolution of intelligence & culture
1:43:43 - Why self driving took so long
1:57:08 - Future of education

Look up Dwarkesh Podcast on YouTube, Apple Podcasts, Spotify, etc. Enjoy!

[Aaron Levie @levie]

Product management for AI agents is easily the wildest form of product management in history.

Typical product management is trying to figure out how to design interfaces and software for people to interact with deterministic systems. The user generally knows all the context to do their work successfully, so it’s generally a matter of nailing the underlying business logic and surrounding UX.

But with AI agents, the user you care about most is the agent, and they don’t know anything by default. They’ll happily run in any direction to perform the task, often without success.

So as a PM (or engineer) you basically spend your time trying to reverse engineer “what would a human need as context to perform this task”, and then figure out how to design systems to get the agent that data in the right sequence, with the right tools, and instructions.

Some of these systems are entire invisible to the human user, but part of the craft is equally how the end-user will interact with the agent to supply this context. Then, it’s often unending trial and error working to eke out incremental points of quality at each stage.

This is especially why people with deep domain expertise, or those that can acquire it quickly, will do extremely well building AI agents. The ability to anticipate the context that the agent would need to be successful is a huge determining factor in how effective the agent will be.

This partly explains why coding agents have worked so well out of the gate; because its builders deeply understand the domain that they’re working to automate. But clearly we’re going to quickly see this same outcome across every field - legal, healthcare, finance, etc. - as context engineering and a new crop of product managers emerge.

[Yongchul Daniel Um @ycdanielum]

@iseongjae299734 nice

[Yongchul Daniel Um @ycdanielum]

@pythaginboots A Liverpool fan here - totally agreed. Reece James without his injuries wouldve taken him and Chelsea places indeed.

[Yongchul Daniel Um @ycdanielum]

@FabrizioRomano Fabrizio lad how long do you sleep haha
Transfer market has not even opened yet hah

[Yongchul Daniel Um @ycdanielum]

@THFCJamie2 Im Korean and no you're not racist.
Koreans have issues with over-idolizing sports stars and I apologize for our country's fans.

[Yongchul Daniel Um @ycdanielum]

@Footywontmiss Messi Henry Lewandowski

[𝗧𝗼𝗽𝗠𝗚☝🏻 🇫🇷⁷ @TopMGSzn]

This is peak South American football 🇺🇾

The striker barely "shoots" towards the goal, an overweight defender fucking dies for no reason, another defender misses the entire ball trying to clear it, then a random dog celebrates the goal in the stands.

[Yongchul Daniel Um @ycdanielum]

@piersmorgan hahahahahahahahahhaahhaahahahahaha