True story: 41 year old VC left his family home to live with 20+ founders half his age, in a house with dorm duties.
Here's why I did it:
SF is back like never before. And @500GlobalVC will be taking the world's best founder talent from across 15 years backing 80 countries to come make it here.
I needed to live it for myself. Stay rooted in what's up and coming. So I can recommend it
If you're up and coming yourself, @theresidency should be your top choice to launch your new life 🔥
Stay tuned: I will be posting more about what goes on inside the homes, what I learned from the new crop of founders, and how to make the most of your stay...
An increasingly ubiquitous average makes a bigger target for deviants. Your insight on designing something 'ugly' is exciting to me. In music, they used to group 'non-pop' into all these other 'alternative' charts. Charts that would birth hiphop, grunge, EDM, emo, etc that cross over and become pop. For design, we will find that raising the bar, like most revolutions, will begin with a little bit of civil disobedience :)
Glad to see the discussion around open v. closed frontier labs. The next area of innovation will be open robotics.
Steel Bot@steelbot_ comes out of stealth with America’s first open humanoid to accelerate the robotics industry. Learn more about the challenge ahead with my conversation with founder @randallmbriggs. https://t.co/lojPXIrkrQ
This week I got to sit down with @TonyW and discuss what I'm working on with @steelbot_. We are building open humanoids that are designed to be the most developer-friendly platform for robotics ever made.
The goal is to empower America's researchers, entrepreneurs, hackers, and hobbyists who want to push the envelope and use general purpose robots in all kinds of applications and industries.
@Brendan_McCord On the same day we left, this popped up at the Cherbourg airport! The experience you created for us is the exact reflective pause needed in the great acceleration. Thank you Brendan @JonathanBi and team!
After extensive usage of Poke by @interaction I’ve decided it’s the most frequent and useful AI service in my life after @claudeai .
However, the more I use it, my experience started to degrade. The underlying issue is something every LLM powered service out there needs to get better at.
It is NOT Memory Bloat.
I call it Memory Fog. It’s different.
I may have given instructions on when to surface what emails - and sometimes, conditions behind when those instructions should be followed.
But how much of my conditions are hidden to Poke (or myself for that matter)? And as I keep piling on new conditions and instructions into Poke’s memory, how should Poke manage MY own memory of what those instructions are?
My own lossy memory = unmanageable expectations = disappointment. Even though the LLM memory architecture itself may have actually got better.
The answer requires memory tuning. Not just for the LLM, but for the human! And the second is harder.
The fix? Default memory tuning from the get go (LLM asking me if I only want emails that are directed to me), proactive memory upgrade suggestions based on usage (eg if it surfaced the emails I ignored), a link to view what’s there, what’s changed, added, changing etc
The best AI products don’t just train the product. They train the human what to expect.
Follow for more UI x AI
As a designer turned 2X exited founder turned VC-builder, I’ve seen UIs from east and west evolve with new waves of tech. With AI, were gonna need need more than a text box. I am passionate about leveraging creative approaches to accelerating AI utility (and adoption).
Will be posting more + hosting a series of talks and meetups in SF on this soon!
Meantime pile on comments on other tips you may have for poke. I’m not an investor there btw, just a fan. Their design approach as a whole is a huge leap.
Day 1 of 500 Silicon Valley Takeover.
@khailee opened with the stat that should reframe every international founder's ambition:
SF raises $2–3B per week in early-stage deals alone. Most countries don't see that in a year.
@albertlee (MyFitnessPal, $475M exit) closed it:
"Moving motivation is very, very difficult. Almost all of our time was spent reducing friction."
"If users succeed, we succeed. That's all that matters."
20 more days to go.
Governments pushing AI literacy are creating a nation of users.
"Users" need Universal Basic Income.
BUILDERS create Universal Basic Innovation.
Literacy isn't enough. Teach people to build.
When orgs use both human and agent, will HR leaders only focus on Human Resources? If they also oversee the orgs agent workforce, what's their new title? And if they don't, report to the Head of Agents, or the other way around?
Didn't expect Jensen Huang to serve me hot green tea. (No, the video isn’t AI-generated).
He spoke about SF, AI, Trump. I asked him about venture capital. But there is more to the story than what he said.
Why he said it, who was there, and what happened during the opera: Embedded in the subtext is a masterful move.
First, the facts:
1. 🐵 He brought his tech superpower friends to see Monkey King - casual!
2. ❤️ He and Lori donate $5M/year to the SF Opera - lovely!
3. 👯 He underscored the role of the arts in the age of AI - important!
4. 🤖 He said “SF is back!” “An AI City” attracting global talent - evident!
5. 🤔 He was one of two people who rang up Trump urging against sending the National Guard into SF —and Trump listened - baller!
6. 🧧 I asked if NVIDIA is going deeper into VC; he gave three reasons (in the video).
And then the subtext:
As I took my seat, I saw Tier 1 VC partners, big-tech CEOs (@sundarpichai included), and decacorn founders do the same. The way everyone settled in felt like a quiet rehearsal for a larger play.
Think about it.
He chose the guests. He chose the opera to back. We weren’t watching “Hangover Part III” :P - We were watcing Monkey King in opera.
Then the curtains parted.
A timeless story about POWER unraveled: The quest for absolute power. To subjugate others or uplift others. To hoard or share power. And most of all, how to use the power you have.
Diane Paulus directed the opera. But Jensen directed the narrative:
As AI accelerates in power, tech titans secure NVIDIA chips to forge AI power -> Venture capital funds AI applications for the masses -> Government leaders battle to secure intelligence as utility for citizens.
Jensen just sat their butts down for a 2.5 hour lecture in full vibrato and monkey makeup - an instruction manual how to use their power like an enlightened being would.
Jensen’s made his message is clear.
Those who continue to accumulate power must remember to empower others with it.
If only more of the leaders he rang up could have joined the night.
( Thanks @chipro@DalianaLiu@nvidia for the invite, and @YengT9 for the inspo)
meet khailee, managing partner at @500GlobalVC
180 seed deals, 7 unicorns, 3 IPO’s
this is a taste of the type of insane talent you could live with by applying to delta
apply now 🔗👇
https://t.co/W88UXB6yg8
@enter_delta
Khailee Ng is a Managing Partner at @500GlobalVC
500 Global is one of the world's most active venture capital firms with over $2.4 billion in assets under management.
He is widely recognized as the architect who transformed 500 from a US-centric seed fund into a truly global platform
Today he gave a presentation to our residents on the physics of fundraising and how to raise your first seed round 🪽 @khailee
Put together by @zeno_fox
Great session with @khailee of @500GlobalVC on the Physics of Fundraising at @theresidency today. Here are my notes, expanded.
Why listen to him
Khailee has three vantage points. First, scrappy founder roots in a capital scarce environment, which created a bias toward cash flow, zero cost user acquisition, and frugality. Second, twelve years as a VC across multiple cycles, sitting with founders through the highs and lows of assembling rounds. Third, he fundraises constantly himself, so he lives the follow up and momentum game.
The core metaphor
Fundraising behaves like physics, with pushes, pulls, and timing. Founders who appear to be “naturals” are usually just fluent in those forces. You can learn that fluency.
1) The ladder strategy
Do not open with your dream firms when your round is empty. Walk up the ladder. Start with easier checks, angels, smaller funds, relevant programs, and non-dilutive money. Build committed capital and social proof, then approach your top targets with momentum in hand. After you land a top name, walk back down the ladder to close fence sitters. Momentum creates urgency, an empty round creates stall.
2) Combination locks
Every investor has a unique code. Pieces of the code include thesis fit, stage preference, fund cycle timing, past portfolio wins and scars, the partner’s angle, even mood and bandwidth. A “no” often reflects their combination, not your worth. Replace victim language like “I got ghosted” with diagnosis, then update your approach. Curiosity beats bitterness.
3) Founder runway beats company runway
If the founder’s energy is depleted, company runway is irrelevant. Protect your tank. Sleep, movement, food, supportive people, a way to reset. If the process drags, you can still win if you preserve energy and perspective. If you burn out, you are out of time regardless of cash in the bank.
4) Simultaneity
Do not pause your company while you raise. Keep shipping, selling, and forming partnerships during the process. Returning to an investor after two months with the same numbers signals dependence on capital. Returning with progress signals you know how to turn dollars into growth. Design your environment to switch contexts on purpose, not by accident. Use your team and advisors as extensions of your brain.
5) Cellular integrity
Say only what you can stand behind fully. Investors have finely tuned spin detectors. Authentic beats theatrical. Be precise about what you have, what you learned, and what the next check unlocks. This builds trust now and reputation later.
6) Signals, not theater
Pedigrees and logos can open a door, they do not build a business. Oversignaling is noise. Share real proof points instead. Conversion and retention that are moving the right way, early payback tests, unit economics trending in the right direction, credible customer anecdotes tied to numbers.
7) Timing and momentum
Categories have seasons. When your category heats up, lean in, compress your process, and consider taking more capital since the window can close. When you are out of season, get even more combination driven and scrappy. Either way, move with intent.
8) Build the fundraising muscle
Reps turn drama into process. Systematize outreach waves, crisp follow ups, learning loops after each call, and weekly progress broadcasts. Treat each interaction as data to refine the deck, proof points, and target list. Conserve energy. High activity is not the goal, effective activity is.
Practical playbook I am taking away
Preparation
• Map investors from easy to aspirational, then plan waves.
• For top targets, write a quick hypothesis of their combination.
• Define a progress plan you will pursue during the raise, independent of capital.
Execution
• Climb the ladder. Collect smaller commits to create social proof and urgency.
• Time outreach in a tight window so people feel momentum, not drift.
• Keep a visible drumbeat of progress that others can reference.
Communication
• Lead with truth. What is working, what is not yet working, what the money unlocks next.
• Offer crisp metrics and milestones, not vanity signals.
• Ask directly for check size, process, timeline, and diligence needs.
Closing
• Once a lead or top logo is secured, loop back to fill the round.
• Protect founder runway with clear boundaries and recovery blocks.
• Close fast. Prolonged raises drain energy and weaken narrative.
Watch outs
• Waiting for money before moving, progress is the strongest signal.
• Personalizing rejection, decode combinations instead.
• Defining success by logos, define it by leverage and learning.
• Slipping into bitterness, it kills momentum and relationships.
Lines that stuck with me
• Fundraising is a ladder, walk up with easy checks, then down to fill.
• Every investor is a combination lock, find the code.
• If the founder is out of energy, the company is out of time.
• Progress during a raise is the clearest proof you will use capital well.
• Authentic composes better than theater, in the room and over the long run.
Grateful for the session, lots to apply immediately.