High Signal AI retweeted
NVIDIA CEO Jensen Huang's answer to the AI race with China, in one sentence: "We should go as fast as we can, but not faster than we should."
Jensen was asked a direct question: Should America be pacing AI development right now, given where it stands against China?
The question ments from Dave (David Sacks), whom Jensen called "incredibly smart," "deep," "pragmatic," and "one of the great American patriots."
@JensenHuang said he wasn't sure of the context of Dave's remarks, but he was confident about where Dave stood:
"I'm fairly certain he wants America to go as fast as we can, but not faster than we should."
Then he repeated the line as his own position:
"We should go as fast as we can, but not faster than we should."
The sentence holds two ideas at once.
"As fast as we can" means no deliberate slowdown and no pacing for its own sake.
"Not faster than we should" means there's a line you don't cross, and for Jensen that line is safety.
What enforces that line? Not goodwill. Consequences:
"If you're compromising safety and you deliver products that are unsafe that put people in harm's way, we have laws of all kinds that come after you."
That's the core of his argument. Companies don't have to be persuaded to build safe AI. Shipping unsafe products already exposes them to the law:
"And so the companies that are developing these technologies, are perfectly incentivized to do the right thing."
His confidence rests on two things: the incentive to build safely and the ability to do it.
"I have every confidence because they're incentivized to do so, they have the capability to do so, that we will deliver, we will create AI, advanced AI in a way that's safe to society."
Jensen was asked a direct question: Should America be pacing AI development right now, given where it stands against China?
The question ments from Dave (David Sacks), whom Jensen called "incredibly smart," "deep," "pragmatic," and "one of the great American patriots."
@JensenHuang said he wasn't sure of the context of Dave's remarks, but he was confident about where Dave stood:
"I'm fairly certain he wants America to go as fast as we can, but not faster than we should."
Then he repeated the line as his own position:
"We should go as fast as we can, but not faster than we should."
The sentence holds two ideas at once.
"As fast as we can" means no deliberate slowdown and no pacing for its own sake.
"Not faster than we should" means there's a line you don't cross, and for Jensen that line is safety.
What enforces that line? Not goodwill. Consequences:
"If you're compromising safety and you deliver products that are unsafe that put people in harm's way, we have laws of all kinds that come after you."
That's the core of his argument. Companies don't have to be persuaded to build safe AI. Shipping unsafe products already exposes them to the law:
"And so the companies that are developing these technologies, are perfectly incentivized to do the right thing."
His confidence rests on two things: the incentive to build safely and the ability to do it.
"I have every confidence because they're incentivized to do so, they have the capability to do so, that we will deliver, we will create AI, advanced AI in a way that's safe to society."
1
5
10
607
High Signal AI retweeted
Meta Chief AI Officer Alexandr Wang on why his college friends, not experienced engineers, were the right people to build Scale AI's early team:
When Alexandr started Scale AI, his first hires didn't come from the San Francisco talent market.
"I mostly hired people I knew from an suggests that was because he could trust them, Alexandr turns it around. It was about whether they could trust him:
"I think more that they could trust me."
Experienced engineers had no reason to. He explains:
"If I went to like a, a 25 year old engineer in San Francisco, I was like, hey, we should, we should work together. I had no credibility."
@alexandr_wang describes how those pitches usually ended:
"I would get coffee with these people and I would say like, yeah, this is what we're working on. Super cool. You should join us. And then they would all just be like, okay, cool, I guess I'm going to go back to my job now."
His college friends were different. They already knew him, and that relationship gave him the credibility that strangers wouldn't:
"So early on, I had no credibility, except for with people I went to college with who we were just like friends and we liked each other. And so I managed to recruit a bunch of them over."
Some of those friends dropped out to join him. Others were seniors who finished school first and then came on board.
"It was a mix. And that was like the early nucleus of the team."
That group of college friends carried Scale AI until the company had something stronger than a coffee pitch to offer: real customers.
"Then we started picking up momentum because we're starting to work with large automotive companies, were starting to work with, you know, these very futuristic autonomous driving companies."
Only then did the team expand beyond that first circle:
"And then as momentum started to pick up, like, you know, we were able to grow and build up the team over time."
The lesson is that your earliest hires are rarely the most experienced people you can find. They are the people who already believe in you. Before you have traction, a relationship can do the work that a track record can't.
If you're building something new with no credibility yet, look first at the people who already know you. They may be the only ones willing to bet on you before anyone else will.
When Alexandr started Scale AI, his first hires didn't come from the San Francisco talent market.
"I mostly hired people I knew from an suggests that was because he could trust them, Alexandr turns it around. It was about whether they could trust him:
"I think more that they could trust me."
Experienced engineers had no reason to. He explains:
"If I went to like a, a 25 year old engineer in San Francisco, I was like, hey, we should, we should work together. I had no credibility."
@alexandr_wang describes how those pitches usually ended:
"I would get coffee with these people and I would say like, yeah, this is what we're working on. Super cool. You should join us. And then they would all just be like, okay, cool, I guess I'm going to go back to my job now."
His college friends were different. They already knew him, and that relationship gave him the credibility that strangers wouldn't:
"So early on, I had no credibility, except for with people I went to college with who we were just like friends and we liked each other. And so I managed to recruit a bunch of them over."
Some of those friends dropped out to join him. Others were seniors who finished school first and then came on board.
"It was a mix. And that was like the early nucleus of the team."
That group of college friends carried Scale AI until the company had something stronger than a coffee pitch to offer: real customers.
"Then we started picking up momentum because we're starting to work with large automotive companies, were starting to work with, you know, these very futuristic autonomous driving companies."
Only then did the team expand beyond that first circle:
"And then as momentum started to pick up, like, you know, we were able to grow and build up the team over time."
The lesson is that your earliest hires are rarely the most experienced people you can find. They are the people who already believe in you. Before you have traction, a relationship can do the work that a track record can't.
If you're building something new with no credibility yet, look first at the people who already know you. They may be the only ones willing to bet on you before anyone else will.
2
5
11
1,102
High Signal AI retweeted
7 unusual websites that can work together to build an $8k/month business
1. Google Trends
1. Google Trends
2
10
18
3,688
High Signal AI retweeted
Lucy Guo on why AI isn't shrinking her workweek: "AI is making people work harder actually, because you have to be awake for your agents to be running."
She pulled a 26-hour day last week because Claude was shipping code and she refused to miss the loop.
She pulled a 26-hour day last week because Claude was shipping code and she refused to miss the loop.
5
7
10
939
High Signal AI retweeted
Tim Cook admits Apple has never once tried to beat competitors to market: “The philosophy has always been to be the best, not the first.”
1
10
27
1,560
High Signal AI retweeted
I found a loophole that's PRINTING ME MONEY 24/7
It's an AI SIDE HUSTLE.
You can earn $3,700 working just a few hours from anywhere in the world.
Comment 'AI' and I will DM you my exclusive guide 100% FREE.
Must follow me so I can send you DM.
FREE for 48 hours only.
It's an AI SIDE HUSTLE.
You can earn $3,700 working just a few hours from anywhere in the world.
Comment 'AI' and I will DM you my exclusive guide 100% FREE.
Must follow me so I can send you DM.
FREE for 48 hours only.
50
7
30
5,989
High Signal AI retweeted
Just my personal opinion here:
Being in real estate and not leveraging SEO or AI Search Optimization in 2026 is like running ads to a business with no landing page.
A huge part of the buying journey now happens before someone ever fills out a form.
Potential clients are ghborhoods, pricing, market conditions and buying decisions across Google, ChatGPT, Gemini and Perplexity.
If you want to see where your site stands across ChatGPT, Claude, Google and broader AI search, start here:
seo-stuff.com/free-audit
Google still relies heavily on relevance, useful content, links and broader site quality, while AI systems can also encounter your business through publications, reviews, directories, videos, forums and other third-party sources.
That overlap is what SEO Stuff was built around:
seo-stuff.com/gold-plan-pack…
For real estate professionals, useful authority plays include guest posts on local housing sites, inclusion in relevant local guides, partnerships with lenders or inspectors, local market commentary and original data around housing trends.
The more credible information that exists about you and your market, the easier it is for search and AI systems to understand where you fit.
Positioning matters too.
“Real estate agent in Los Angeles” tells people very little.
“Buyer’s agent specializing in first-time homebuyers in Beverly Hills, Brentwood and Santa Monica” is far more useful.
The more specific you are about who you help, where you work and what you specialize in, the easier it is to match your business to a high-intent search.
Then look at the website itself.
Your site should act as the main source of truth about you and your business.
Keep it fast, mobile-friendly and easy to crawl, use clean URLs and build logical internal links between your services, neighborhoods, buyer and seller resources and contact pages.
At minimum, you want a clear homepage, a strong About page, service or listings pages, detailed neighborhood guides and an easy way to contact or book with you.
Content structure matters too.
AI systems frequently reuse specific passages rather than entire pages, so make the useful information easy to find.
Instead of burying the answer, use headings such as “How much are closing costs in Pasadena?” or “What should first-time buyers know about Burbank?”
Answer directly, then expand with useful context, examples, numbers and local expertise.
Comparison content works well too.
Think “Condo vs. single-family home in Los Angeles,” “Pasadena vs. South Pasadena for families” or “Renting vs. buying in LA in 2026.”
Those are real decisions buyers are trying to make.
Keyword research should follow the same logic.
Look at the full questions buyers and sellers ask before contacting somebody.
Examples might include “how to buy a home in Pasadena with bad credit,” “best neighborhoods in LA for families,” “how much does it cost to sell a house in California” and “what should I know before buying an investment property in Malibu.”
Those searches may have less volume, but the person behind them often has a very specific problem.
You should also stop letting your best social proof disappear inside social feeds.
Turn useful Reels, TikToks, client questions and market commentary into permanent assets such as blog posts, YouTube videos, neighborhood guides, FAQs and newsletter archives.
Video is especially useful for real estate because the product itself is visual.
Create neighborhood walkthroughs, market updates, buyer education and customer stories, then publish clear titles, descriptions and transcripts so the information is easy to understand and discover.
Branded search matters too.
When someone searches your name, you want your own site, profiles, reviews and useful content to make it obvious who you are, where you work and why someone might hire you.
That means maintaining accurate profiles, earning real reviews and publishing content that reinforces your actual expertise.
Finally, make it easy for traffic to turn into a lead.
Use clear calls to action, booking options, buyer or seller guides, email follow-up and simple contact forms.
Someone researching a neighborhood at 10pm should not need to hunt around the site to figure out how to speak with you.
The bigger point is pretty simple.
Real estate buyers and sellers increasingly research before they reach out.
If your competitors are the ones answering their questions across Google and AI search, those competitors are getting introduced earlier in the decision.
Build local authority, publish useful market information, own the questions your customers ask and make your expertise easy to understand.
Or let SEO Stuff handle the content and authority side for you:
seo-stuff.com
And if you want to see where your site stands across ChatGPT, Claude, Google and broader AI search, start here:
seo-stuff.com/free-audit
Being in real estate and not leveraging SEO or AI Search Optimization in 2026 is like running ads to a business with no landing page.
A huge part of the buying journey now happens before someone ever fills out a form.
Potential clients are ghborhoods, pricing, market conditions and buying decisions across Google, ChatGPT, Gemini and Perplexity.
If you want to see where your site stands across ChatGPT, Claude, Google and broader AI search, start here:
seo-stuff.com/free-audit
Google still relies heavily on relevance, useful content, links and broader site quality, while AI systems can also encounter your business through publications, reviews, directories, videos, forums and other third-party sources.
That overlap is what SEO Stuff was built around:
seo-stuff.com/gold-plan-pack…
For real estate professionals, useful authority plays include guest posts on local housing sites, inclusion in relevant local guides, partnerships with lenders or inspectors, local market commentary and original data around housing trends.
The more credible information that exists about you and your market, the easier it is for search and AI systems to understand where you fit.
Positioning matters too.
“Real estate agent in Los Angeles” tells people very little.
“Buyer’s agent specializing in first-time homebuyers in Beverly Hills, Brentwood and Santa Monica” is far more useful.
The more specific you are about who you help, where you work and what you specialize in, the easier it is to match your business to a high-intent search.
Then look at the website itself.
Your site should act as the main source of truth about you and your business.
Keep it fast, mobile-friendly and easy to crawl, use clean URLs and build logical internal links between your services, neighborhoods, buyer and seller resources and contact pages.
At minimum, you want a clear homepage, a strong About page, service or listings pages, detailed neighborhood guides and an easy way to contact or book with you.
Content structure matters too.
AI systems frequently reuse specific passages rather than entire pages, so make the useful information easy to find.
Instead of burying the answer, use headings such as “How much are closing costs in Pasadena?” or “What should first-time buyers know about Burbank?”
Answer directly, then expand with useful context, examples, numbers and local expertise.
Comparison content works well too.
Think “Condo vs. single-family home in Los Angeles,” “Pasadena vs. South Pasadena for families” or “Renting vs. buying in LA in 2026.”
Those are real decisions buyers are trying to make.
Keyword research should follow the same logic.
Look at the full questions buyers and sellers ask before contacting somebody.
Examples might include “how to buy a home in Pasadena with bad credit,” “best neighborhoods in LA for families,” “how much does it cost to sell a house in California” and “what should I know before buying an investment property in Malibu.”
Those searches may have less volume, but the person behind them often has a very specific problem.
You should also stop letting your best social proof disappear inside social feeds.
Turn useful Reels, TikToks, client questions and market commentary into permanent assets such as blog posts, YouTube videos, neighborhood guides, FAQs and newsletter archives.
Video is especially useful for real estate because the product itself is visual.
Create neighborhood walkthroughs, market updates, buyer education and customer stories, then publish clear titles, descriptions and transcripts so the information is easy to understand and discover.
Branded search matters too.
When someone searches your name, you want your own site, profiles, reviews and useful content to make it obvious who you are, where you work and why someone might hire you.
That means maintaining accurate profiles, earning real reviews and publishing content that reinforces your actual expertise.
Finally, make it easy for traffic to turn into a lead.
Use clear calls to action, booking options, buyer or seller guides, email follow-up and simple contact forms.
Someone researching a neighborhood at 10pm should not need to hunt around the site to figure out how to speak with you.
The bigger point is pretty simple.
Real estate buyers and sellers increasingly research before they reach out.
If your competitors are the ones answering their questions across Google and AI search, those competitors are getting introduced earlier in the decision.
Build local authority, publish useful market information, own the questions your customers ask and make your expertise easy to understand.
Or let SEO Stuff handle the content and authority side for you:
seo-stuff.com
And if you want to see where your site stands across ChatGPT, Claude, Google and broader AI search, start here:
seo-stuff.com/free-audit
Almost 73% of the websites cited across ChatGPT, Claude, Gemini and Perplexity appeared in just ONE of them.
Only 2.2% were cited by all four.
So if you are checking ChatGPT to see whether your company is “showing up in AI search,” you may be looking at a very incomplete ives us a useful look at just how differently the major AI platforms research and cite the web.
Let’s go through it.
And if you want to see where your business stands across ChatGPT, Claude, Google AI, Perplexity, Grok and broader AI search, start here. It’s free:
seo-stuff.com/free-audit
Prefer ran the same 80 questions through ChatGPT, Perplexity, Gemini and Claude.
Each question was tested three times on each platform, and that produced 960 AI answers and 1,329 different cited websites.
The results were pretty interesting.
966 of those websites appeared in only one AI engine.
205 appeared in two.
129 appeared in three.
Just 29 appeared across all four.
Basically, there really was no universal “AI search result.”
Different systems can research the exact same topic and pull information from very different parts of the web.
SEO Stuff's Done-for-you setup was created with this in mind:
seo-stuff.com/gold-plan-pack…
The number of sources they cited was dramatically different too.
Perplexity averaged 19.48 sources per answer.
Gemini averaged 8.64.
Claude averaged 4.58.
ChatGPT averaged just 3.05.
Across the 240 answers generated by each platform, Perplexity cited 748 different websites, Gemini 580, Claude 333 and ChatGPT 218.
That is why asking “How do I rank in AI?” is not as informative as it sounds.
The questions that make a bit more sense (and will help you more long-term) are:
How often do I appear in ChatGPT?
What does Claude know about my company?
Which sources does Perplexity trust in my category?
What is Gemini surfacing for the questions my buyers ask?
There was another important finding, by the way.
The researchers asked the same questions three times within seconds of each other, yet the cited websites still changed.
Average overlap between repeat runs was:
96% for Perplexity.
68% for Claude.
37% for ChatGPT.
35% for Gemini.
Same question.
That matters enormously for measurement because if you ask ChatGPT one question once and do not see your company, that does not necessarily mean you never appear.
And if you do see your company once, that does not mean you reliably own that answer either.
You need repeated measurements across multiple prompts and multiple platforms, otherwise you can end up building an entire AI strategy around a screenshot.
So what about sources?
Well, Reddit appeared in 207 of the 960 answers, making it the most frequently cited site in the study.
Claude cited Reddit zero times.
Gemini cited YouTube in 113 answers.
ChatGPT, Claude and Perplexity cited YouTube zero times in this dataset.
Perplexity cited LinkedIn in 73 answers.
Claude cited it twice.
ChatGPT and Gemini did not cite it at all.
Again, the study focused on 80 questions about AI search, GEO and AEO. Healthcare, finance, ecommerce or local-service prompts could produce a very different source mix, but the larger point is difficult to ignore.
Different AI systems have different retrieval behavior, different source ecosystems and different levels of citation consistency.
That said, the biggest authoritative sources traveled much better across platforms.
Among the 25 most-cited sites, 23 appeared across three or four engines.
Five of the top 10 appeared across all four.
So the fragmentation was much heavier in the long tail, and that may actually be the most useful takeaway for businesses.
You want enough content breadth to appear across many relevant questions, but you also want enough authority that multiple systems repeatedly trust your website and the third-party sources validating your company.
That is where traditional SEO and AI Search Optimization start overlapping again.
Build useful content around the questions buyers actually ask.
Make your company, products and category easy to understand.
Earn credible third-party mentions and contextual backlinks.
Create comparisons, alternatives, use cases and decision-stage content.
That is exactly why SEO Stuff approaches AI visibility across multiple platforms rather than treating it as a ChatGPT-only exercise:
seo-stuff.com
A company can perform well in ChatGPT and poorly in Claude.
If your company only measures one AI platform, you may be missing most of the picture.
And if you want to see where your business stands across Google, ChatGPT, Claude and the major AI search platforms, start here:
seo-stuff.com/free-audit
11
28
39
4,540
High Signal AI retweeted
Dumb 9-to-5 workers:
• Wait 40 years to retire.
• Depend on annual salary increases.
• Spend weekends on beer & ordering pizza.
Smart 9-to-5 workers:
• Wait 40 years to retire.
• Depend on annual salary increases.
• Spend weekends on beer & ordering pizza.
Smart 9-to-5 workers:
3
18
45
15,781
High Signal AI retweeted
Skip Coursera, Udemy, and expensive playbooks.
MIT has 10 amazing free AI courses:
Sharing a great course compilation from Anisha Jain.
1. AI 101
Clear introduction to core AI ideas.
Ideal if you’re starting from zero and want the big picture first.
🔗 ocw.mit.edu/courses/res-6-…
2. Introduction to Deep Learning
Hands-on bootcamp covering neural networks and modern deep learning foundations.
🔗 introtodeeplearning.com
3. Artificial Intelligence
Covers problem-solving, knowledge representation, and machine learning fundamentals.
🔗 ocw.mit.edu/courses/6-034-…
4. Introduction to Machine Learning
Beginner-friendly overview of ML principles and essential algorithms you’ll actually use.
🔗 openlearninglibrary.mit.edu/courses/course…
5. How to AI (Almost) Anything
Explore how AI applies to music, art, perception, and creative systems.
🔗 ocw.mit.edu/courses/mas-s6…
6. Understanding the World Through Data
Use data analysis and ML techniques to interpret real-world problems.
🔗 edx.org/learn/data-sci…
7. Artificial Intelligence in K-12 Education
Foundations of AI technology and how it can shape learning environments.
🔗 ocw.mit.edu/courses/6-s062…
8. Introduction to Algorithms
A key prerequisite for building efficient AI systems and writing optimized code.
🔗 ocw.mit.edu/courses/6-006-…
9. Foundation Models and Generative AI
Deep dive into modern AI systems (like ChatGPT) and how large models work.
🔗 ocw.mit.edu/courses/6-s087…
10. Minds and Machines
Exploring reality, consciousness, AI, and more.
🔗 learn.mit.edu/courses/course…
Focus on one AI track at a time.
Dedicate 30–45 minutes daily and practice as you go.
Apply your learning in small projects or real tasks to grow quickly.
♻️ Share this so more people can access these free resources.
——
📌 Get my top free course compilations:
1) Follow me.
2) Subscribe to my free newsletter at WildCapital.co.
You’ll receive them directly in your welcome email.
MIT has 10 amazing free AI courses:
Sharing a great course compilation from Anisha Jain.
1. AI 101
Clear introduction to core AI ideas.
Ideal if you’re starting from zero and want the big picture first.
🔗 ocw.mit.edu/courses/res-6-…
2. Introduction to Deep Learning
Hands-on bootcamp covering neural networks and modern deep learning foundations.
🔗 introtodeeplearning.com
3. Artificial Intelligence
Covers problem-solving, knowledge representation, and machine learning fundamentals.
🔗 ocw.mit.edu/courses/6-034-…
4. Introduction to Machine Learning
Beginner-friendly overview of ML principles and essential algorithms you’ll actually use.
🔗 openlearninglibrary.mit.edu/courses/course…
5. How to AI (Almost) Anything
Explore how AI applies to music, art, perception, and creative systems.
🔗 ocw.mit.edu/courses/mas-s6…
6. Understanding the World Through Data
Use data analysis and ML techniques to interpret real-world problems.
🔗 edx.org/learn/data-sci…
7. Artificial Intelligence in K-12 Education
Foundations of AI technology and how it can shape learning environments.
🔗 ocw.mit.edu/courses/6-s062…
8. Introduction to Algorithms
A key prerequisite for building efficient AI systems and writing optimized code.
🔗 ocw.mit.edu/courses/6-006-…
9. Foundation Models and Generative AI
Deep dive into modern AI systems (like ChatGPT) and how large models work.
🔗 ocw.mit.edu/courses/6-s087…
10. Minds and Machines
Exploring reality, consciousness, AI, and more.
🔗 learn.mit.edu/courses/course…
Focus on one AI track at a time.
Dedicate 30–45 minutes daily and practice as you go.
Apply your learning in small projects or real tasks to grow quickly.
♻️ Share this so more people can access these free resources.
——
📌 Get my top free course compilations:
1) Follow me.
2) Subscribe to my free newsletter at WildCapital.co.
You’ll receive them directly in your welcome email.
9
90
415
15,772
High Signal AI retweeted
Almost 73% of the websites cited across ChatGPT, Claude, Gemini and Perplexity appeared in just ONE of them.
Only 2.2% were cited by all four.
So if you are checking ChatGPT to see whether your company is “showing up in AI search,” you may be looking at a very incomplete ives us a useful look at just how differently the major AI platforms research and cite the web.
Let’s go through it.
And if you want to see where your business stands across ChatGPT, Claude, Google AI, Perplexity, Grok and broader AI search, start here. It’s free:
seo-stuff.com/free-audit
Prefer ran the same 80 questions through ChatGPT, Perplexity, Gemini and Claude.
Each question was tested three times on each platform, and that produced 960 AI answers and 1,329 different cited websites.
The results were pretty interesting.
966 of those websites appeared in only one AI engine.
205 appeared in two.
129 appeared in three.
Just 29 appeared across all four.
Basically, there really was no universal “AI search result.”
Different systems can research the exact same topic and pull information from very different parts of the web.
SEO Stuff's Done-for-you setup was created with this in mind:
seo-stuff.com/gold-plan-pack…
The number of sources they cited was dramatically different too.
Perplexity averaged 19.48 sources per answer.
Gemini averaged 8.64.
Claude averaged 4.58.
ChatGPT averaged just 3.05.
Across the 240 answers generated by each platform, Perplexity cited 748 different websites, Gemini 580, Claude 333 and ChatGPT 218.
That is why asking “How do I rank in AI?” is not as informative as it sounds.
The questions that make a bit more sense (and will help you more long-term) are:
How often do I appear in ChatGPT?
What does Claude know about my company?
Which sources does Perplexity trust in my category?
What is Gemini surfacing for the questions my buyers ask?
There was another important finding, by the way.
The researchers asked the same questions three times within seconds of each other, yet the cited websites still changed.
Average overlap between repeat runs was:
96% for Perplexity.
68% for Claude.
37% for ChatGPT.
35% for Gemini.
Same question.
That matters enormously for measurement because if you ask ChatGPT one question once and do not see your company, that does not necessarily mean you never appear.
And if you do see your company once, that does not mean you reliably own that answer either.
You need repeated measurements across multiple prompts and multiple platforms, otherwise you can end up building an entire AI strategy around a screenshot.
So what about sources?
Well, Reddit appeared in 207 of the 960 answers, making it the most frequently cited site in the study.
Claude cited Reddit zero times.
Gemini cited YouTube in 113 answers.
ChatGPT, Claude and Perplexity cited YouTube zero times in this dataset.
Perplexity cited LinkedIn in 73 answers.
Claude cited it twice.
ChatGPT and Gemini did not cite it at all.
Again, the study focused on 80 questions about AI search, GEO and AEO. Healthcare, finance, ecommerce or local-service prompts could produce a very different source mix, but the larger point is difficult to ignore.
Different AI systems have different retrieval behavior, different source ecosystems and different levels of citation consistency.
That said, the biggest authoritative sources traveled much better across platforms.
Among the 25 most-cited sites, 23 appeared across three or four engines.
Five of the top 10 appeared across all four.
So the fragmentation was much heavier in the long tail, and that may actually be the most useful takeaway for businesses.
You want enough content breadth to appear across many relevant questions, but you also want enough authority that multiple systems repeatedly trust your website and the third-party sources validating your company.
That is where traditional SEO and AI Search Optimization start overlapping again.
Build useful content around the questions buyers actually ask.
Make your company, products and category easy to understand.
Earn credible third-party mentions and contextual backlinks.
Create comparisons, alternatives, use cases and decision-stage content.
That is exactly why SEO Stuff approaches AI visibility across multiple platforms rather than treating it as a ChatGPT-only exercise:
seo-stuff.com
A company can perform well in ChatGPT and poorly in Claude.
If your company only measures one AI platform, you may be missing most of the picture.
And if you want to see where your business stands across Google, ChatGPT, Claude and the major AI search platforms, start here:
seo-stuff.com/free-audit
Only 2.2% were cited by all four.
So if you are checking ChatGPT to see whether your company is “showing up in AI search,” you may be looking at a very incomplete ives us a useful look at just how differently the major AI platforms research and cite the web.
Let’s go through it.
And if you want to see where your business stands across ChatGPT, Claude, Google AI, Perplexity, Grok and broader AI search, start here. It’s free:
seo-stuff.com/free-audit
Prefer ran the same 80 questions through ChatGPT, Perplexity, Gemini and Claude.
Each question was tested three times on each platform, and that produced 960 AI answers and 1,329 different cited websites.
The results were pretty interesting.
966 of those websites appeared in only one AI engine.
205 appeared in two.
129 appeared in three.
Just 29 appeared across all four.
Basically, there really was no universal “AI search result.”
Different systems can research the exact same topic and pull information from very different parts of the web.
SEO Stuff's Done-for-you setup was created with this in mind:
seo-stuff.com/gold-plan-pack…
The number of sources they cited was dramatically different too.
Perplexity averaged 19.48 sources per answer.
Gemini averaged 8.64.
Claude averaged 4.58.
ChatGPT averaged just 3.05.
Across the 240 answers generated by each platform, Perplexity cited 748 different websites, Gemini 580, Claude 333 and ChatGPT 218.
That is why asking “How do I rank in AI?” is not as informative as it sounds.
The questions that make a bit more sense (and will help you more long-term) are:
How often do I appear in ChatGPT?
What does Claude know about my company?
Which sources does Perplexity trust in my category?
What is Gemini surfacing for the questions my buyers ask?
There was another important finding, by the way.
The researchers asked the same questions three times within seconds of each other, yet the cited websites still changed.
Average overlap between repeat runs was:
96% for Perplexity.
68% for Claude.
37% for ChatGPT.
35% for Gemini.
Same question.
That matters enormously for measurement because if you ask ChatGPT one question once and do not see your company, that does not necessarily mean you never appear.
And if you do see your company once, that does not mean you reliably own that answer either.
You need repeated measurements across multiple prompts and multiple platforms, otherwise you can end up building an entire AI strategy around a screenshot.
So what about sources?
Well, Reddit appeared in 207 of the 960 answers, making it the most frequently cited site in the study.
Claude cited Reddit zero times.
Gemini cited YouTube in 113 answers.
ChatGPT, Claude and Perplexity cited YouTube zero times in this dataset.
Perplexity cited LinkedIn in 73 answers.
Claude cited it twice.
ChatGPT and Gemini did not cite it at all.
Again, the study focused on 80 questions about AI search, GEO and AEO. Healthcare, finance, ecommerce or local-service prompts could produce a very different source mix, but the larger point is difficult to ignore.
Different AI systems have different retrieval behavior, different source ecosystems and different levels of citation consistency.
That said, the biggest authoritative sources traveled much better across platforms.
Among the 25 most-cited sites, 23 appeared across three or four engines.
Five of the top 10 appeared across all four.
So the fragmentation was much heavier in the long tail, and that may actually be the most useful takeaway for businesses.
You want enough content breadth to appear across many relevant questions, but you also want enough authority that multiple systems repeatedly trust your website and the third-party sources validating your company.
That is where traditional SEO and AI Search Optimization start overlapping again.
Build useful content around the questions buyers actually ask.
Make your company, products and category easy to understand.
Earn credible third-party mentions and contextual backlinks.
Create comparisons, alternatives, use cases and decision-stage content.
That is exactly why SEO Stuff approaches AI visibility across multiple platforms rather than treating it as a ChatGPT-only exercise:
seo-stuff.com
A company can perform well in ChatGPT and poorly in Claude.
If your company only measures one AI platform, you may be missing most of the picture.
And if you want to see where your business stands across Google, ChatGPT, Claude and the major AI search platforms, start here:
seo-stuff.com/free-audit
If you want ChatGPT to recommend your business when someone asks who they should buy from, there is a pretty straightforward opportunity hiding in plain sight.
New data across 289,105 cited URLs across ChatGPT, Google AI Mode, AI Overviews and Perplexity offers a unique window
Let’s go through it.
And if you want to see whether ChatGPT, Claude, Google AI, Perplexity, Grok and broader AI search are already recommending your business, start here (it's free):
seo-stuff.com/free-audit
Surfer analyzed nearly 1000 prompts across 12 industries, generating 26,573 AI calls and 289,105 cited URLs over seven days.
Then they asked a fairly simple question:
When an AI system recommends several brands, is there a relationship between how prominently each brand appears and how many of the cited sources mention that brand?
Spoiler alert: there was.
Brands mentioned across a larger share of the sources being used by the AI tended to appear higher in the response.
Obviously that does NOT mean Surfer proved that brand mentions cause higher AI rankings, but the relationship is strong enough that businesses working on AI visibility should probably pay attention.
Think about how this can work.
Someone asks:
“What are the best accounting platforms for a small nonprofit?”
The AI retrieves information from several sources.
Maybe multiple articles mention Company A, compare it with competitors and describe its use cases.
Company B appears once.
Company A now has a much larger body of information surrounding it across the source set.
Your website matters, but the wider web matters too, and the results became even more interesting when Surfer broke them down by platform.
ChatGPT showed the strongest relationship between brand presence across cited sources and recommendation position among the systems tested.
It also cited substantially more sources in this dataset: 84.7% more than Google AI Mode and 129.7% more than Google AI Overviews.
So if ChatGPT is working from a larger source pool, there are potentially more places where your brand can either appear or be absent.
This is why treating AI Search Optimization as something that happens entirely on your own website does not make much sense.
Your website is one source and the system may be looking at many others.
And if your competitors repeatedly appear across those sources while your business does not, they are giving the AI a much broader information footprint to work with.
That is exactly why SEO Stuff combines content with authority:
seo-stuff.com/gold-plan-pack…
The content side helps your own website explain who you are, what you do, who you serve, which problems you solve, how you compare and what buyers should know.
The authority side creates additional third-party references around the company.
There is an important distinction here though, as this study measured BRAND MENTIONS, not backlinks.
A backlink can still matter for traditional SEO, but Surfer measured whether the cited page actually mentioned the brand.
That means a good editorial placement can potentially do two different jobs.
It can provide a legitimate backlink that supports traditional search authority, while also putting your brand, category and context onto a third-party page an AI system may encounter.
The research found something particularly interesting about blogs too.
Brand-owned blogs represented 16.2% of cited sources and third-party blogs another 12.5%, meaning blogs accounted for 28.7% of the cited source set.
But brand mentions appearing in blog content showed the strongest correlation with recommendation position of any content type Surfer analyzed.
That matters because blogs are one of the areas businesses can actually influence.
You can publish useful content, create original research, contribute expert commentary, earn editorial coverage and get included in relevant industry articles.
But one of the most valuable things you can do is identify the pages AI systems already cite for the commercial questions your customers ask.
Do not just think:
“I need more backlinks.”
Think:
“What does ChatGPT cite when someone asks who the best company in my category is?”
“What sources does Google AI Mode use?”
“Which comparison pages keep showing up?”
“Which of those pages mention my competitors but not me?”
That creates a much more useful target list than simply sorting websites by Domain Rating.
If you're starting from scratch today, here's how I'd play it:
Start with the commercial questions that matter, then look at the sources AI systems actually use when answering them.
Which brands keep appearing? Which websites keep getting cited? Where are competitors mentioned? Where are you missing?
Then work on both sides.
Build your own content so AI systems have accurate, useful first-party information about your company, and build a credible third-party footprint so your website is not the only place on the internet associating your brand with the category.
That is increasingly what AI Search Optimization looks like.
Useful content, commercial topic coverage, clear brand information, relevant third-party mentions and contextual authority across the web.
And if you want to see where your business currently stands across ChatGPT, Claude, Google AI, Perplexity, Grok and broader AI search, start here:
seo-stuff.com/free-audit
30
32
73
12,234
Nobody takes a break between jobs. You finish Friday, start Monday, and show up to the hardest work of your career already worn out.
Serval is paying new engineering hires to take a week off before they start. Up to $10k, planned with a concierge. Referrals get $30k each way on top.
They raised $150M at $3B last week and are building code generation for the parts of the enterprise that never got it. IT, HR, finance, and legal.
Serval is paying new engineering hires to take a week off before they start. Up to $10k, planned with a concierge. Referrals get $30k each way on top.
They raised $150M at $3B last week and are building code generation for the parts of the enterprise that never got it. IT, HR, finance, and legal.
Starting today, every engineer who joins Serval begins with a weeklong, $10,000 paid vacation. We’re building the next generational company. We need generational talent to see it through.
Our mission has always been to empower people across every team to turn their understanding ng, trustworthy automations. Making that function inside the largest enterprises takes world-class engineering.
That’s what you’d be joining us to do. There are foundational decisions still to make, and the engineers making them will shape the future of how big we can build this company.
We want you to arrive rested and ready to do the best work of your life. So your first week is a chance to get away.
And if a @getserval employee refers you, both you and that employee receive $30,000, on top of the vacation.
Come build with us: serval.com/careers
0
0
0
481
Your data costs don’t scale with how much data you have. They scale with how many things read it.
Every dashboard, model and agent gets its own pipeline on the same events. That was fine when adding a consumer meant a human asking for one, a few times a year.
Agents cost nothing to create and never stop reading. The pipeline-per-consumer model doesn’t survive that.
Eon Yukon does it in one pass instead, as the events land. Quarter of Spark’s compute, seventh of Flink’s, benchmark’s public.
Every dashboard, model and agent gets its own pipeline on the same events. That was fine when adding a consumer meant a human asking for one, a few times a year.
Agents cost nothing to create and never stop reading. The pipeline-per-consumer model doesn’t survive that.
Eon Yukon does it in one pass instead, as the events land. Quarter of Spark’s compute, seventh of Flink’s, benchmark’s public.
What I'm seeing across the industry right now is that the biggest bottleneck to AI adoption isn't model capabilities. It's data infrastructure.
Today, we're expanding the Eon Data Foundation with the launch of Eon Yukon, a new serverless data engine purposely built to change resh, contextualized data to act.
Specialized models need a continuous feed of data and feedback to improve. Both are still served by a stack, an architecture, and methods built for dashboards:
-Spark and Flink clusters sized for peak, plus compaction jobs to keep tables usable.
-A pipeline and a copy for every consumer, staged bronze to silver to gold.
-Hand-written jobs, wired together by an orchestrator and kept alive by specialists.
The result: shipping data is complicated, risky, and expensive. While coding agents ship apps in hours, they rarely build or run data pipelines. The data is too volatile, and a mistake costs too much.
With Yukon, data moves from ingestion to processing to serving in one continuous flow, without the infrastructure and pipelines traditionally required to connect them.
Events become compacted, sorted, deduplicated Iceberg tables within seconds.
Real-time aggregations run inline, with fully managed state. Agents and applications can access fresh data with sub-millisecond latency through a Redis-compatible API or through our SQL and spark compatible query engine.
Beyond moving and serving data, Yukon makes it continuously understandable - Live column-level statistics, accessible through MCP, give agents visibility into the data. Built-in monitoring, sampling, and data minimization help teams maintain control.
At 1M events/sec and equal freshness, Yukon used 4x less compute than Spark and 7x less than Flink. The benchmark is public: run it yourself with Eon’s ingest-bench.
This is another major step toward our vision of an Autonomous Data Foundation for the AI era, built to both protect enterprise data and unlock its full potential.
1
4
7
2,693
The uncanny valley in AI avatars was never really about rendering. It's about latency.
A face that looks perfect but replies 400 ms late reads as wrong, because conversation runs on timing. The pause before an answer carries meaning. Miss it and the thing feels dead regardless are.
Which is why most AI video is pre-rendered. Generate the clip, and play the clip. Nobody is doing it live at conversational speed because the compute doesn't allow it.
Atmee v2 goes at that constraint. One image, 10 seconds of audio, and a live character you can video-call up to 50x faster and 100x cheaper than leading video models.
Realtime isn't a faster version of generation. It's the thing that makes it feel like someone's there.
A face that looks perfect but replies 400 ms late reads as wrong, because conversation runs on timing. The pause before an answer carries meaning. Miss it and the thing feels dead regardless are.
Which is why most AI video is pre-rendered. Generate the clip, and play the clip. Nobody is doing it live at conversational speed because the compute doesn't allow it.
Atmee v2 goes at that constraint. One image, 10 seconds of audio, and a live character you can video-call up to 50x faster and 100x cheaper than leading video models.
Realtime isn't a faster version of generation. It's the thing that makes it feel like someone's there.
After Google acquired my AR startup, I started thinking about the next interface, this time for AI. Text and voice work fine, just like DOS did before the Mac. But nothing’s more intuitive than face to face.
After two years of research, today we’re launching Atmee v2: a that makes live video agents feel truly life-like.
• Any character from 1 image + 10s of audio
• Contextual expressions, gestures, even interacting with the environment
• Live video calls, voice only, or video generation
At its core, Atmee v2 is a video foundation model, it can render anything, but it’s built to excel at characters, which is what video models like Seedance 2.5 and Wan 3.0 are mostly used for. In comparison:
Atmee v2
• Up to 50x faster
• 10-100x cheaper
Build tutors, sales coaches, support agents, companions, game characters, even whole living worlds. The interface after text and voice starts today.
1
2
7
2,024
A context window is a desk, not a memory. Everything’s in reach, and nothing survives the session.
So the fix has been a bigger desk. More tokens, same information re-read from scratch every turn, at cost.
Memory works the other way. It decides what’s worth keeping and throws rain does that outside the model. In their internal evals, there are up to 22.6x fewer tokens per question.
Scaling context was a way of avoiding the harder question of what to forget.
So the fix has been a bigger desk. More tokens, same information re-read from scratch every turn, at cost.
Memory works the other way. It decides what’s worth keeping and throws rain does that outside the model. In their internal evals, there are up to 22.6x fewer tokens per question.
Scaling context was a way of avoiding the harder question of what to forget.
We decoupled memory from the model.
Living Brain is one persistent brain for ChatGPT, Claude, Cursor, Codex and your own agents. Write in one, retrieve in any.
92.2% on LoCoMo. 464ms median retrieval. Up to 22.6x fewer tokens per question.
LLMs have parameters. Living Brain has plasticity.
1
2
5
1,662
Voice agents are built, tested, and graded on transcripts. That’s not one bad product, it’s the industry’s whole measurement layer.
A transcript keeps what was said and throws away how. The sarcastic “oh, perfect” scores as positive sentiment. The third, louder, "ask for a human" reads as the word “representative” three times. The agent passes every eval and loses the customer.
Human panels catch all of it, which is why they’re the gold standard and why nobody runs them per checkpoint.
Hume’s Decision API makes that judgment an API call. Audio in, your question, probabilities back, trained on 10M+ human judgments.
Voice got good faster than anyone could measure it.
A transcript keeps what was said and throws away how. The sarcastic “oh, perfect” scores as positive sentiment. The third, louder, "ask for a human" reads as the word “representative” three times. The agent passes every eval and loses the customer.
Human panels catch all of it, which is why they’re the gold standard and why nobody runs them per checkpoint.
Hume’s Decision API makes that judgment an API call. Audio in, your question, probabilities back, trained on 10M+ human judgments.
Voice got good faster than anyone could measure it.
Introducing the Hume Audio Decision API.
Ask anything: How many speakers? Is anyone laughing? Is the caller frustrated?
What used to take 10+ models is now one call.
Trained on 10M+ human judgments, so it answers the way human listeners would.
Realtime, or across hours of cess below.
1
2
7
2,579
Most AI for architecture stops at the test fit, where everything is still a shape and mistakes are cheap.
This one models the bolts on a glass facade from the manufacturer’s spec sheet.
snaptrude.com
This one models the bolts on a glass facade from the manufacturer’s spec sheet.
snaptrude.com
Today we're launching @snaptrude Agent, the most advanced agent for building design.
It does real design work with you, from a site to a construction-ready LOD 400 model with structure, MEP and drawings. Then can go back to your existing software with your families intact.
t the company on AI. Honestly the insight came from software engineering. It wasn't long ago that autocomplete felt like a big deal, and now engineers direct agents that write a lot of the code. I think the same shift is coming to how we design buildings, and it's already started.
Most designers I meet aren't short on skill, they're short on time. Take the call every design team dreads, the client cuts the budget by 30%. Today that's a bad month for everyone. With the agent you're the director. You decide where the cost comes out, and it updates the layout, coordinates the structure and MEP, and redoes the drawings. So when design exploration gets this cheap, we go from 3 options to 300, and more options means better buildings.
Since August our users have gone through more than 140 billion tokens, and firms like Jacobs, Studio+, Kajima, VMDO, Dekker and hundreds of others use Snaptrude today. In our early access, more than 2,000 designers have run over 25,000 tasks through the agent since September.
We also built a benchmark for this, AECBench, because in AEC an almost correct answer is still wrong. It gives AI the same brief an architect would get, has its design in a real BIM model, and scores the result against the brief and against how architects review design. We're publishing the first results today, link below.
Thank you @prashanthp at @Accel and Shub and the team at @foundamental for backing us when this was still a bet, and the Snaptrude team, who built every bit of this. Really proud of you all.
Thanks to our investors, @scottbelsky, @andreasklinger, @thilokonzok, @immad, @CindyXBi, @DanilaVladi, @jmwass, rahulchaudhary, @adamecourt, Zach Ginsburg, Alex duff, Paul grossinger, really appreciate all the help along the way.
0
1
7
2,110
Liquid fine-tuned a 2B model to spot defects in drone photos of wind turbine blades. About 3,000 public photos, a few hours on a single A100.
Before the fine-tune it caught fewer than 10% of defects. After, 67% classification accuracy. A frontier model on the same held-out set stayed at the bottom, often locating the defect correctly and then labeling it wrong.
Generality and expertise aren’t the same axis. The frontier race optimizes the first one. Almost every real deployment needs the second.
LFM3 is built for that end. Six 2B-class models on one backbone: text, vision, audio, omni, decision, and open weights, each fine-tunable on a single GPU.
A model that knows everything is rarely the best model for the one thing you’re doing.
Before the fine-tune it caught fewer than 10% of defects. After, 67% classification accuracy. A frontier model on the same held-out set stayed at the bottom, often locating the defect correctly and then labeling it wrong.
Generality and expertise aren’t the same axis. The frontier race optimizes the first one. Almost every real deployment needs the second.
LFM3 is built for that end. Six 2B-class models on one backbone: text, vision, audio, omni, decision, and open weights, each fine-tunable on a single GPU.
A model that knows everything is rarely the best model for the one thing you’re doing.
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence.
> 6 model instances, across text, audio, vision, and omni modalities, gendels.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding
> 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages.
> 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language.
> 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input.
> fine-tune on 1 GPU.
> a new tri-hybrid architecture
> architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒
Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device.
1/ 🧵
2
2
9
2,677
AI doing AI research has been the thing that happens next for about forty years. It’s in the papers, it’s on every panel, and it’s always a few years out.
The reason it matters is compounding. A better research system finds better discoveries, which improve the research system, which finds better discoveries. Most work in AI is linear. That loop isn’t.
Recursive built one and pointed it at a model. The Eureka Machine ideates, implements, validates, then uses the results to pick what to try next.
It contributed across the stack" architecture, low-bit training, inference kernels, speculative decoding.
What came out is Flint. A 40B open-weight coding model, best in its class on agentic coding benchmarks, 125+ tokens/sec on a MacBook at short contexts, Apache 2.0.
Small team, under four months, from scratch.
The reason it matters is compounding. A better research system finds better discoveries, which improve the research system, which finds better discoveries. Most work in AI is linear. That loop isn’t.
Recursive built one and pointed it at a model. The Eureka Machine ideates, implements, validates, then uses the results to pick what to try next.
It contributed across the stack" architecture, low-bit training, inference kernels, speculative decoding.
What came out is Flint. A 40B open-weight coding model, best in its class on agentic coding benchmarks, 125+ tokens/sec on a MacBook at short contexts, Apache 2.0.
Small team, under four months, from scratch.
Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine.
• Best-in-class fast local coding
• Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark
• Built from hs
Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.
1
0
7
614
The app store assumes what you need already exists, built by someone else, for a problem close enough to yours.
Most of what people actually want fails that test. A tracker for one odd habit. A way to settle one recurring argument in one group chat. Too specific for anyone to built.
When software can be generated on demand, that constraint goes away. The question stops being whether an app exists and becomes whether you can describe it.
Wabi puts that in a messenger… Personal agents, group chats, and apps made for whatever you’re working on.
The app store was always a workaround for software being too expensive to make for one person.
Most of what people actually want fails that test. A tracker for one odd habit. A way to settle one recurring argument in one group chat. Too specific for anyone to built.
When software can be generated on demand, that constraint goes away. The question stops being whether an app exists and becomes whether you can describe it.
Wabi puts that in a messenger… Personal agents, group chats, and apps made for whatever you’re working on.
The app store was always a workaround for software being too expensive to make for one person.
Wabi 2.0 is live!
personal agents are one person, one chat. but the parts of life that matter run on more than one person
Wabi is an agent you use together. you talk to it, it does stuff, builds the apps you need, and anyone can join the thread - your partner, your best friend, pchat
school changes pickup time → Wabi updates the family calendar, everyone sees it.
a show you’d both love → Wabi finds a night that works, gets the tickets.
the trip everyone keeps talking about → Wabi picks the dates, builds the itinerary, books it
go start a thread with your people 👇
2
8
12
2,342