andrew chen(@andrewchen)

welcome @seeamshahidnoor to the Speedrun team!! Joining us from Scale AI 💪🏼

andrew chen (@andrewchen) — Wed, 22 Jul 2026 16:42:09 GMT

welcome @seeamshahidnoor to the Speedrun team!! Joining us from Scale AI 💪🏼

Seeam Shahid Noor@seeamshahidnoor

BIG Update: I’m joining @a16z as an Investing Partner on the a16z @speedrun team! 🎉

I’ll be focusing on early-stage investing as part of speedrun, where we invest up to $1M in exceptional founders building great companies, & give them unfair advantages to succeed. If you are or know a founder who wants to win, I'd love to meet you! I'm especially excited about compelling novel ideas in AI dev tools, agent RL & evals, data, & infra in general.

For those who have known me over the years, you know how much joy I get in supporting those around me. From working countless nights to support my brilliant teammates at @scale_AI, to spending thousands of hours in my leisure time making free videos for millions of students in Bangladesh, I have given my all to support those doing their life's most important work. That has always been my personal mission.

So, when I met the speedrun team, a team that works relentlessly to make early bets & support founders from the very beginning of their journey, I knew this was the dream team for me. I am thrilled to join this shared mission with @andrewchen, @Tocelot, @tkexpress11, @emilybenn12, @kenanhsaleh, @far33d, @marcussegal, @ndrewlee, & the rest of the team! It's time to build.

🔗 View Quoted Tweet

💬6🔄3❤️45👀12796📊8

⚡ Powered by xgo.ing

for folks thinking about a role change this summer, we want to find you a job at a startup -- we’ve ...

andrew chen (@andrewchen) — Tue, 14 Jul 2026 17:40:48 GMT

for folks thinking about a role change this summer, we want to find you a job at a startup -- we’ve invested in >350 companies so far via a16z speedrun.

I’ve spent time with all of the founders: I’m betting (literally) that they’re the next Cursor, Kalshi, Decagon, Databricks, etc.

their jobs are never listed - they tend to hire via referral or from our intros via the speedrun talent team

want to work with them? Let us know here: bit.ly/4fe2B4V

if the team sees a match, they’ll reach out to intro you to the founders.

💬40🔄45❤️426👀863551📊169

⚡ Powered by xgo.ing

new thing gets invented “Ban it!” “Install communism!” “No more new products!” “There’s a bubble!” ...

andrew chen (@andrewchen) — Tue, 14 Jul 2026 01:51:37 GMT

new thing gets invented

“Ban it!”
“Install communism!”
“No more new products!”
“There’s a bubble!”
“Think of the children!”
“We need a committee!”

(And no I’m not just talking about AI)

💬40🔄25❤️311👀55360📊73

⚡ Powered by xgo.ing

before AI: Drop shipping bros - making Shopify stores after: SaaS shipping bros - making niche vert...

andrew chen (@andrewchen) — Mon, 13 Jul 2026 19:28:12 GMT

before AI:
Drop shipping bros - making Shopify stores

after:
SaaS shipping bros - making niche vertical AI apps

Only half kidding here, but how many times do I click on someone's X profile after seeing a post showing a cool little app demo, and see that they have like 10 other app projects all making $10k/month lol

💬57🔄5❤️170👀33187📊70

⚡ Powered by xgo.ing

the 1-autist unicorn theory: if AI removes the need for people to build, it removes the need for pe...

andrew chen (@andrewchen) — Mon, 13 Jul 2026 19:11:22 GMT

the 1-autist unicorn theory:

if AI removes the need for people to build, it removes the need for people skills...

this means that ultra high IQ / ultra low EQ autist incels become the most important and valuable people on the planet, going forward

not just the 1-person unicorn company theory, but the 1-autist unicorn theory. QED

💬115🔄11❤️318👀47517📊146

⚡ Powered by xgo.ing

the interview alpha from asking "which social media platform do you use the most?" is going to go to...

andrew chen (@andrewchen) — Mon, 13 Jul 2026 02:18:41 GMT

the interview alpha from asking "which social media platform do you use the most?" is going to go to zero when the X feed is ruled by photo carousels, isn't it?

cc @signulll

💬11🔄0❤️48👀18444📊14

⚡ Powered by xgo.ing

very common convo in tech rn: Startup: “We want to launch can you intro us to [anti-tech news org]?...

andrew chen (@andrewchen) — Mon, 13 Jul 2026 00:10:01 GMT

very common convo in tech rn:

Startup: “We want to launch can you intro us to [anti-tech news org]?”

“uh probably don’t want to do that, they’ll probably ignore or write a hit piece, not much in the middle”

Startup: “ok but we really want to”

… a few weeks later …

“They fucked us over” or…
“They ignored us”
Not much in the middle unfort

dar@radbackwards

I gave WIRED the exclusive on our hands launch, and they wrote a really weird article about how we are sexualizing robotics… 

wired.com/story/the-1x-n…

I felt pretty betrayed because that’s not what they told me they were writing about not is that what I’ve ever been about… actually I stand for quite the opposite… But I’ve come to find a lot of dishonesty and malice in the journalism community so I wasn’t surprised. 

This is what I sent the author… I’m only sharing this because I hope it encourages journalists to resist the click bait trap and tell truly awesome stories because I for one don’t believe journalism is dead— I think it’s just starting and just needs to evolve past the weird corner of the internet where data driven optimization turns everything into smooth brained shocking brain rot bullshit. 

The technological revolution we are going through should inspire a journalism renaissance. Not let it fall into further decay. There is so much brilliance at play in the world and the stories should be told!

My note: 

“[author name redacted], it was nice talking to you, but I wanted to let you know that I didn’t enjoy your article at all. 

I understand the need to be inflammatory because that seems to be the only thing that gets clicks these days but that doesnt mean you shouldn’t recognize when something special is in front of you. 

I trusted our PR team in saying we should offer you the exclusive on what is one of the most important technological developments in the history of Mankind and I deeply regret it. 

Good luck with the rest of your writing career. 

-Dar Sleeper”

🔗 View Quoted Tweet

💬39🔄8❤️311👀101782📊57

⚡ Powered by xgo.ing

"640k should be enough for anybody"

andrew chen (@andrewchen) — Sun, 12 Jul 2026 23:27:03 GMT

"640k should be enough for anybody"

💬1🔄0❤️12👀4218📊2

⚡ Powered by xgo.ing

playing with local AI models and doing some research and came across a fun extrapolation: "consumer...

andrew chen (@andrewchen) — Sun, 12 Jul 2026 23:25:12 GMT

playing with local AI models and doing some research and came across a fun extrapolation:

"consumer grade graphics cards will be running Fable-equivalent models by 2029"

the argument:

1) been said that LLMs are really a lossy compression process for squeezing humanity's knowledge down into a chat box. Roughly you could say the entire internet/books/media/whatever is about ~200-300T tokens which gets "compressed" down to ~1T params of frontier models, so imagine a 400:1 or so compression in bytes. (We have a lot more video and world data, we'll discuss that later...)

2) People further quantize the models down (seemingly at decent quality) from 16-bit to 4-bit (NVFP4 ftw) and it seems pretty good. So that's more like 1,600:1 or so. Research suggests models only store ~2 bits of knowledge per parameter anyway — the weights were mostly air, which is why quantization works at all.

3) furthermore, we seem to be getting better at this compression, whether it's with MoE, pruning, etc. The Qwen 27B dense model now is equivalent to higher parameter models from a few years ago. 1,600:1 today and 1,600:1 in 5 years will have completely different results

4) I was looking to see if there's a Moore's Law thing happening here, and it's been measured: the "Densing Law" found capability-per-parameter doubles every ~3.3 months. Not sure how well it'll hold, but it says somethign like: "Every 3 months, the size of model needed to represent humanity's knowledge drops by half."

5) Not sure this holds though bc presumably, there's some kind of asymptote. Won't compress down to zero, the same way that modern image/video compression has theoretic limits too

5) Video is a zillion frames, almost zero semantic density. 99.99% of every frame is stuff a physics prior already predicts. World models will post compression ratios in the millions-to-one. But, nevertheless, there will be a ton of new facts seen simply by observing, that was never written down. But even without dealing with all this, today's text-oriented frontier LLMs are already pretty amazing

6) So the crazy idea here is that ultimately irreducible kernel representing humanity's knowledge might compress down to... tens of GB? May be small enough to fit onto a consumer grade GPU. Today, a consumer grade GPU for playing video games might have 32GB of memory on it, but the 27B parameter model that thing can run is getting smarter and smarter each year. Will the equivalent of Fable be able to run on a high-end consumer GPU in a few years? 

This sounds crazy, but GPT-4 was rumored ~1.8T params, needed a rack of A100s, cost tens of millions to train. Today's open weight models with 27B parameters can do that with better data, distillation, and architecture.

That's the big question. Densing Law says "consumer GPUs might be running Fable-equivalent in 2028" and it might be possible, and even though that would be 100x? Seems nuts

💬27🔄13❤️124👀21000📊50

⚡ Powered by xgo.ing

This is what I want to say: Been playing w GLM 5.2 for the past week and it’s legit. If I can get t...

andrew chen (@andrewchen) — Sun, 12 Jul 2026 22:21:21 GMT

This is what I want to say:

Been playing w GLM 5.2 for the past week and it’s legit. If I can get this in a box sitting on my desk, I’m not sure I need much else

But this is probably what’s actually true:

Can imagine a guy running on a Pentium PC for the first time and being like, wow, I don’t need anything else. But actually we invent all sorts of cool new things we can use compute. So each frontier/adjacent LLM is cool but we’ll always want to use the next, better one

💬35🔄9❤️134👀38279📊45

⚡ Powered by xgo.ing

hour 1: I am a 10x engineer hour 4: What have I done and why is it 4,000 lines

andrew chen (@andrewchen) — Fri, 10 Jul 2026 23:58:36 GMT

hour 1:
I am a 10x engineer

hour 4:
What have I done and why is it 4,000 lines

💬37🔄7❤️445👀37201📊72

⚡ Powered by xgo.ing

some of you know that i've been tinkering with local AI models on my homelab setup. It's fun but her...

andrew chen (@andrewchen) — Fri, 10 Jul 2026 21:56:47 GMT

some of you know that i've been tinkering with local AI models on my homelab setup. It's fun but here's the reality today:

- nothing you can run on consumer hardware is even close to what you can buy with a monthly subscription. The gap is ~18 years when you compare frontier to frontier, but you're actually comparing frontier cloud LLMs (1 trillion parameter models) to what you can run locally which might be 30-100 billion. This is a hobby!

- best hardware off the shelf is probably a new macbook pro with lots of RAM or a gaming PC. Or if you want, a DGX Spark from Nvidia. A new M5 Macbook Pro with 32/48GB is probably the minimum spec for most. Most people's machines don't have enough memory though, and small models are not that good

- this is because you need to run, say, a quantized version of qwen 35b and above to get good/coherent answers back from your prompts. If you run a 9b parameter model bc that's what fits, you won't like the results. Slow and bad for most use cases

- you need 30-50 tok/s for it to be usable from a UX perspective. This is determined mostly by memory bandwidth which is determined by the type of GPU. So again, macbook pro or a gaming PC

- normie prompts work pretty well now, the kind of stuff you'd plug into google. ("what is the population..." "what's the best..."). You don't want to code using local models - there's a huge huge gap

however even with all these drawbacks, you learn a ton! And it's fun to experiment. I have 2x DGX sparks and a 5090 eGPU in my homelab and it's great. And you can always run GLM or whatever open weight frontier model on openrouter as a backup also. Def interesting to tinker with context window sizes, new tech like DSpark/etc.

Over the next few years, a good number of white collar workers will run "good enough" models on their macbook pros issued at work and that'll unlock a lot of interesting new use cases. So when you get a new MBP from work, ask for more memory!

💬33🔄3❤️96👀22406📊45

⚡ Powered by xgo.ing

underrated use of agentic coding harnesses: IT for fixing your hermes/openclaw integrations IT for ...

andrew chen (@andrewchen) — Fri, 10 Jul 2026 21:39:43 GMT

underrated use of agentic coding harnesses:

IT for fixing your hermes/openclaw integrations
IT for securing your homelab
IT for debugging your local AI setup
etc

Must have saved me 100s of hours at this point... so many weird bugs that you encounter by trying out random github codebases lol

💬13🔄1❤️36👀9033📊18

⚡ Powered by xgo.ing

Agentic coding has the most epic subscription+upsell pricing model: code code code → run out of tok...

andrew chen (@andrewchen) — Fri, 10 Jul 2026 20:58:29 GMT

Agentic coding has the most epic subscription+upsell pricing model:

code code code → run out of tokens → buy more → rinse and repeat

it's a subscription with a gas pedal. the meter never stops

💬29🔄1❤️71👀13495📊31

⚡ Powered by xgo.ing

Also I’m pumped about this upcoming movie

andrew chen (@andrewchen) — Fri, 10 Jul 2026 18:37:08 GMT

Also I’m pumped about this upcoming movie

Cinema Tweets@CinemaTweets1

FIRST LOOK: The Brink of War is an upcoming political thriller about President Reagan's efforts to negotiate nuclear disarmament with Soviet leader Mikhail Gorbachev. Jeff Daniels plays Ronald Reagan & Jared Harris plays Mikhail Gorbachev while J.K. Simmons plays George Shultz. In other words, the cast is loaded. I’m partial to any political thriller, it’s a genre I love. This is added to my @letterboxd watchlist.

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬0🔄0❤️6👀3482📊1

⚡ Powered by xgo.ing

small % of prompts drive all the value large % of prompts drive all the retention

andrew chen (@andrewchen) — Thu, 09 Jul 2026 23:16:44 GMT

small % of prompts drive all the value
large % of prompts drive all the retention

andrew chen@andrewchen

Pepsi challenge for LLMs

Contrarian view during a week of huge new model launches:

All of us do a lot of “normie prompts” - these are use cases which are really like Google searches (“what’s the name of..” “is it true that…” “what’s the best…”). These are a very high % of total prompts- maybe not in terms of value creation (like code gen or the frontiers of math/science we’re going to) but it’s ubiquitous

If you plugged these LLM prompts into the various frontier models could they tell the difference on the quality of output? I think not. We’d all fail in a blind taste test I think, as the models are now “good enough”

we’re already at the point of diminishing returns in terms of what LLMs return back for a large % of use cases. And there’s implications:

1) open source models will constitute the majority of LLM queries. Open weight models lag by 18-24 months but adding to the question above, could you tell the difference on non-frontier local AI models that can run on modern Mac hardware? I’ve been doing exactly this with models like Qwen 27b dense and honestly they’re great for the normie prompts. There’s a huge incentive for NVIDIA, apple, and maybe even handset manufacturers like Samsung/etc to host open weight AI as an add on to just get you to buy their software

2) AI pricing heads to zero. And we’ll see free and ad-supported AI will be a thing in the consumer market, and open weight models are part of the story here too. Seems like we are <12-18 months to being able to just have ad supported AI particularly for developing markets and segments where the monthly fee doesn’t make sense. Monthly/metered might just be a thing in B2B use cases

3) once quality differences even out the competitive dimension shifts to other factors. Privacy, interconnectivity, free, bundling. The other idea here is that the moat becomes the wrapper (err we call them harnesses now? lol) and the product built around the LLM. 

4) of course premium/frontier models will continue to exist. As long as there are big differences outside of the normie prompts, then you’ll hire one LLM over another for world generation, coding, science, labor replacement/augmentation etc. Just saying I’m not sure we’ll need frontier models for 90%+ of consumer use cases

I think the prevalence of benchmarking in the launch of new AI models is in agreement with this. This week I tried Grok 4.5 and Fable for some coding experiments and you need to really spend time to pick up the differences. So we use benchmarks to point out what’s not so obvious

Some of us will remember when computers were all measured in megahertz and megabytes, and the PC industry compared itself that way. Over time, that gave way to design, power efficiency, etc. Today we’re benchmarking and calculating cost per token and so on. It’s about to evolve, I think

🔗 View Quoted Tweet

💬6🔄1❤️15👀10977📊6

⚡ Powered by xgo.ing

Pepsi challenge for LLMs Contrarian view during a week of huge new model launches: All of us do a ...

andrew chen (@andrewchen) — Thu, 09 Jul 2026 20:25:01 GMT

Pepsi challenge for LLMs

Contrarian view during a week of huge new model launches:

All of us do a lot of “normie prompts” - these are use cases which are really like Google searches (“what’s the name of..” “is it true that…” “what’s the best…”). These are a very high % of total prompts- maybe not in terms of value creation (like code gen or the frontiers of math/science we’re going to) but it’s ubiquitous

If you plugged these LLM prompts into the various frontier models could they tell the difference on the quality of output? I think not. We’d all fail in a blind taste test I think, as the models are now “good enough”

we’re already at the point of diminishing returns in terms of what LLMs return back for a large % of use cases. And there’s implications:

1) open source models will constitute the majority of LLM queries. Open weight models lag by 18-24 months but adding to the question above, could you tell the difference on non-frontier local AI models that can run on modern Mac hardware? I’ve been doing exactly this with models like Qwen 27b dense and honestly they’re great for the normie prompts. There’s a huge incentive for NVIDIA, apple, and maybe even handset manufacturers like Samsung/etc to host open weight AI as an add on to just get you to buy their software

2) AI pricing heads to zero. And we’ll see free and ad-supported AI will be a thing in the consumer market, and open weight models are part of the story here too. Seems like we are <12-18 months to being able to just have ad supported AI particularly for developing markets and segments where the monthly fee doesn’t make sense. Monthly/metered might just be a thing in B2B use cases

3) once quality differences even out the competitive dimension shifts to other factors. Privacy, interconnectivity, free, bundling. The other idea here is that the moat becomes the wrapper (err we call them harnesses now? lol) and the product built around the LLM. 

4) of course premium/frontier models will continue to exist. As long as there are big differences outside of the normie prompts, then you’ll hire one LLM over another for world generation, coding, science, labor replacement/augmentation etc. Just saying I’m not sure we’ll need frontier models for 90%+ of consumer use cases

I think the prevalence of benchmarking in the launch of new AI models is in agreement with this. This week I tried Grok 4.5 and Fable for some coding experiments and you need to really spend time to pick up the differences. So we use benchmarks to point out what’s not so obvious

Some of us will remember when computers were all measured in megahertz and megabytes, and the PC industry compared itself that way. Over time, that gave way to design, power efficiency, etc. Today we’re benchmarking and calculating cost per token and so on. It’s about to evolve, I think

💬24🔄3❤️38👀19069📊27

⚡ Powered by xgo.ing

Don’t forget to check out the F117 Nighthawk outside 😎🇺🇸 I mentioned recently I’m reading “Skunk...

andrew chen (@andrewchen) — Wed, 08 Jul 2026 21:18:02 GMT

Don’t forget to check out the F117 Nighthawk outside 😎🇺🇸

I mentioned recently I’m reading “Skunk Works” which has a long section on the plane’s development. Really remarkable engineering

💬3🔄0❤️15👀4719📊4

⚡ Powered by xgo.ing

visited the Reagan presidential library in SoCal and it’s 👌👌👌 - lots on the Cold War, including ...

andrew chen (@andrewchen) — Wed, 08 Jul 2026 21:17:47 GMT

visited the Reagan presidential library in SoCal and it’s 👌👌👌

- lots on the Cold War, including a section of the Berlin War
- special 250 anniversary exhibit w an British East India tea box from the Boston Tea Party
- Air Force One (Boeing 707, used by 7 presidents from Nixon to Reagan)
- of course a ton from Reagan’s childhood and Hollywood career to his political life

💬6🔄1❤️78👀14218📊13

⚡ Powered by xgo.ing

lol all the best movies have this critic:audience ratings ratio

andrew chen (@andrewchen) — Mon, 06 Jul 2026 21:58:09 GMT

lol all the best movies have this critic:audience ratings ratio

💬6🔄0❤️46👀11862📊9

⚡ Powered by xgo.ing

I grew up in Seattle. Sad what’s happened there

andrew chen (@andrewchen) — Sun, 05 Jul 2026 22:11:04 GMT

I grew up in Seattle. Sad what’s happened there

Brian Heywood - That Damn Mormon@bkheywood

This is almost the perfect Seattle Times article

Warning TLDR 
but shorter than reading the actual article

1) it “identifies” a major problem years after the problem has become obvious

36% vacancy of office buildings in Seattle!

2) the problem identified was caused almost entirely by policies from politicians the Seattle Times endorsed but this is not acknowledged so there is no quote here to pass on

3) the article then uses the super effective “some people say” to set up their ultimate goal of shushing away the true problem

“Some commentators have blamed the downtown office apocalypse on Seattle’s taxes, antibusiness rhetoric and perceptions of public safety. .”

4) but then quickly pivots to the ST-go-to-patented move  —blaming the victim for the crime.   In this case it blames Tech.  See, it is techs fault that Seattle is 36% vacant.  

“The bigger culprit, though, is the tech sector. Its astonishing decadelong push for office space, and equally astonishing slowdown, left downtown Seattle with a gap between supply and demand that will be very difficult to bridge. “

5) Finally it offers a useless solution that ignores the root cause and tries to put a bandage on the excessive blood flow.

In this case the solution appears to be to backfill/conversion of all the empty offices spaces that actually employ people with apartments …I am guessing for the unemployed???

Well, Brian, with that impressive list, why did you say you only give it an “almost” perfect score? 
Well, quite frankly, they did not find a way to pull in global warming… Although… they did get very close with a zombie winter warning…. 

“For some market watchers, conversion’s slow start is simply one more indicator that the office sector is in for an extended winter”

So so close… where is the ST paid propaganda Climate Lab editors when you need them?

🔗 View Quoted Tweet

💬5🔄7❤️92👀30315📊13

⚡ Powered by xgo.ing

Some of my favorite small towns: - Durango, CO - Bisbee, AZ - Moab, UT - Lone Pine, CA - Ojai, CA - ...

andrew chen (@andrewchen) — Sat, 04 Jul 2026 20:49:48 GMT

Some of my favorite small towns:
- Durango, CO
- Bisbee, AZ
- Moab, UT
- Lone Pine, CA
- Ojai, CA
- Oatman, AZ
- Nevada City, CA
- Panguitch UT
- Estes Park, CO
- Whitefish MT
- Williams AZ

💬6🔄0❤️24👀4663📊10

⚡ Powered by xgo.ing

In celebration of America’s 250 years, a book to add to your reading list - some wonderful stories a...

andrew chen (@andrewchen) — Sat, 04 Jul 2026 19:15:00 GMT

In celebration of America’s 250 years, a book to add to your reading list - some wonderful stories about the creation of:
- F 117 nighthawk
- U2 spy plane
- SR71 blackbird
… and more

Wonderful read! H/t @jasonrosenthal

💬23🔄12❤️262👀78130📊70

⚡ Powered by xgo.ing

during Covid, instead of being barricaded at home in SF, I got into a RV and started working (thanks...

andrew chen (@andrewchen) — Sat, 04 Jul 2026 17:48:37 GMT

during Covid, instead of being barricaded at home in SF, I got into a RV and started working (thanks Starlink!) from the road from states I’d prev never been to. I saw our country, its vastness and beauty, its small towns, historic monuments and parks

It made me appreciate this great country even more! 🇺🇸🇺🇸🇺🇸🇺🇸🇺🇸!!!!

Highly recommend, especially for those who spend most of their time in front of computer screens and dense/urban cities with all of their pluses and minuses

Some spots to bookmark that you’ve maybe never considered:
- minuteman missile national historic site / titan missile museum
- rodeo in Cody WY
- world museum of mining in Butte MT
- Watch Oppenheimer then go to Los Alamos History Museum
- mighty 5 loop in southern Utah and stay at Ulum for glamping
- million dollar highway in Colorado
- ghost towns in bodie, Jerome, Virginia city, etc
- drive the 395 in the eastern sierras and visit the Alabama hills
- slab city
- get gas and a tshirt at Amboy

I could go on and on but this is a good start! There’s nothing like an RV to make you appreciate the USA!!!!

💬17🔄9❤️225👀26258📊45

⚡ Powered by xgo.ing

🇺🇸🇺🇸🇺🇸!!!!!!

andrew chen (@andrewchen) — Fri, 03 Jul 2026 16:15:26 GMT

🇺🇸🇺🇸🇺🇸!!!!!!

Your browser does not support the video tag.

🔗 View on Twitter

💬10🔄6❤️69👀23467📊20

⚡ Powered by xgo.ing

wen cyber taxi edition model Y long? (We can never be satisfied)

andrew chen (@andrewchen) — Thu, 02 Jul 2026 20:07:37 GMT

wen cyber taxi edition model Y long?
(We can never be satisfied)

Tesla@Tesla

Introducing Model Y Long Wheelbase – now available in the US & Puerto Rico

A 3-row, 6-seat configuration that brings exceptional interior space with ample headroom & legroom for all passengers

0-60 in 4.4 seconds
325 miles of range

– Front row: Heated/ventilated seats w/ powered thigh cushion
– Second row: Heated/ventilated captain seats w/ powered armrests & one‑touch fold
– Third row: Heated seats w/ power recline, one‑touch fold & child seat anchors

–  Improved airflow, increased efficiency & more range

– 89 cu ft of trunk space: with 6 passengers, trunk still fits a 28" and 20" suitcase each, plus frunk holds an additional 20" suitcase. Oversized items like snowboards and bikes also fit easily

– Engineered for safety w/ seat belts & side air bag optimized for second & third row

– Upgraded acoustic glass & suspension to minimize road noise

– Adaptive damping for a smooth, stable ride

– Staggered tires for enhanced grip

– Larger tailgate for better rear visibility & bigger windows overall to deliver great views from every seat

– 16" first row & 8” second row touchscreens + 19-speaker immersive Tesla Audio

– Upgraded 50W wireless charging pads w/ active cooling & charging ports for all other seats

– FSD Supervised & integrated Grok AI

tesla.com/modely/design#…

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬2🔄1❤️31👀16684📊4

⚡ Powered by xgo.ing

Related chart. Let’s make people rich This is why the tech above will unlock nannies, drivers, surr...

andrew chen (@andrewchen) — Thu, 02 Jul 2026 18:35:46 GMT

Related chart. Let’s make people rich

This is why the tech above will unlock nannies, drivers, surrogacy and other things that will help make having kids much easier than it is now

💬4🔄0❤️6👀2792📊5

⚡ Powered by xgo.ing

I am meh about declining fertility trends bc there’s so much tech coming to encourage fertility and ...

andrew chen (@andrewchen) — Thu, 02 Jul 2026 00:14:18 GMT

I am meh about declining fertility trends bc there’s so much tech coming to encourage fertility and population growth:

- robonannies / robotic household staff
- embryos made from stem cells / skin cells
- longevity tech, particularly of reproductive windows
- GLP1 / Turkish hair transplants / other lookmaxxing tech
- automated housing construction
- artificial wombs
- autonomous driving
- tradwife influencers and inevitable Taylor swift babies

… what else?

Plus AI tutors, genetic embryo screening, etc to make sure we raise better kids too!

💬58🔄9❤️169👀35639📊77

⚡ Powered by xgo.ing

Incredible use of AI and FOIA. Congrats @rmcentush (Shar pei pug, Aphrodite)

andrew chen (@andrewchen) — Tue, 30 Jun 2026 20:30:53 GMT

Incredible use of AI and FOIA. Congrats @rmcentush 

(Shar pei pug, Aphrodite)

Ryan McEntush@rmcentush

We FOIA’d every San Francisco dog license from the last decade: 

82,070 license events, 51,379 dogs, 13,525 names, 412 breeds, 26 neighborhoods.

Then we turned it into a data story: The Dogs of San Francisco.

🔗 View Quoted Tweet

💬6🔄1❤️28👀27920📊8

⚡ Powered by xgo.ing

software used to be the cheap part of labor - tools Now labor is the expensive fallback for software...

andrew chen (@andrewchen) — Mon, 29 Jun 2026 18:05:57 GMT

software used to be the cheap part of labor - tools
Now labor is the expensive fallback for software - the non-verifiable/non-automatable part

💬19🔄2❤️40👀10059📊20

⚡ Powered by xgo.ing

my god it’s beautiful

andrew chen (@andrewchen) — Mon, 29 Jun 2026 04:02:01 GMT

my god it’s beautiful

kerman kohli@kermankohli

why see the leaning tower of pisa when you can see the leaning power of dgx sparks

🔗 View Quoted Tweet

💬4🔄2❤️33👀17200📊6

⚡ Powered by xgo.ing

2002: I'll google it 2015: I'll uber over 2026: I'll GPT it (???). I'll claude it (???) Why haven't...

andrew chen (@andrewchen) — Fri, 26 Jun 2026 00:19:09 GMT

I'll google it
I'll uber over
I'll GPT it (???). I'll claude it (???)

Why haven't we picked the dominant verb for AI yet?

💬1174🔄21❤️3416👀563013📊1272

⚡ Powered by xgo.ing

whoops - i missed a shadow IT step, but maybe that's workplace normies

andrew chen (@andrewchen) — Thu, 25 Jun 2026 23:11:19 GMT

whoops - i missed a shadow IT step, but maybe that's workplace normies

💬2🔄0❤️3👀4534📊2

⚡ Powered by xgo.ing

current AI model supply chain: AI researchers → US AI labs → distilled Chinese models → SF startups...

andrew chen (@andrewchen) — Thu, 25 Jun 2026 22:27:02 GMT

current AI model supply chain:

AI researchers → US AI labs → distilled Chinese models → SF startups → prosumers → workplace normies → Fortune 500 procurement

Am I wrong?

💬24🔄4❤️70👀20303📊31

⚡ Powered by xgo.ing

The always insightful @davidu 👇🏼

andrew chen (@andrewchen) — Thu, 25 Jun 2026 19:39:32 GMT

The always insightful @davidu 👇🏼

Suffiyan Malik@suffiyanmalikk

I got to sit down with David Ulevitch (@davidu) for E31 of We The Builders (@WeThe_Builders). He is a General Partner at @a16z where he cofounded the $1.776 billion American Dynamism practice with @KTmBoyle about 4 years ago. Since then, it has taken over the timeline and become the language many use to describe frontier, deep or hard tech. Friend of the show @lochhead calls this Category Design, where you don’t necessarily have to be the first to become a category leader, you need to frame the problem and do it better than everyone else.

We talk about investing strategy, founding story of American Dynamism and what made it a sticky brand, what he has learnt about how the world works from his anthropology background, surveillance concerns with public safety technologies, his best mentor who is not @pmarca and more. 

American Dynamism focuses on investing in areas such as public safety, aerospace, logistics, national security, defense, manufacturing, industrialization, logistics. Some of the companies in their portfolio include @anduriltech , Amca, @AppliedInt , Apex Space, @AstroMechanica , @Astranis , @basepowerco , @CAForever , @Castelion , @HadrianInc , @NorthwoodSpace , @RadiantNuclear  and of course @SpaceX. Founders of many of these will be featured in upcoming episodes or are friends of the show.

David was previously the founder and CEO of OpenDNS, a cloud-delivered security service that was acquired by Cisco in 2015 for $635 million. While at @Cisco , David was Senior Vice President and General Manager of Cisco Security — a $2.4 billion annual revenue business with more than 5,000 team members — where he oversaw the company’s global cybersecurity strategy, product portfolio, and business (via A16Z).

Timestamps:

00 - Introduction
57 - Defining American Dynamism
43 - Project Maven and Rebuilding the Arsenal of Democracy
24 - Competing with General Catalyst for Anduril
49 - Category creation and defining the market
57 - The marketing engine of “The American Dream”
20 - Replicating Silicon Valley
22 - Robot Dogs and General Catalyst
47 - 1776 Million: Building the American Dynamism Practice
59 - Why Hardware Outperforms Software Over Time
38 - Coupling Defense, Energy, and Space
28 - Investing in Public Safety
01 - Investment Committees are “stupid”
39 - Long-term partners over short-term bets
23 - Policy, Regulations, and navigating relationships in DC
11 - How regulations are slowing us down
54 - The Industrial Supply Chain Divergence
29 - Global Maritime Dominance and Expanding TAM
39 - Venture-Backable Bets in Critical Industrial Infrastructure
00:24 - Applying Anthropology: Study of People, Culture, and Political Power
04:08 - What Drives Innovation Across Nations?
11:49 - Regional Talent and State Incentives
15:48 - Moving from Benevolent Dictator to Minority Investor
20:42 - Life is Checkers, Not Chess: Agency and Knowing When to Sell
26:57 - OpenDNS and the Pivot to Enterprise: Letting Markets Dictate Your Revenue
30:13 - The High Bar for Talent: Grittiness, Grind, and Magnetic Leadership
33:09 - Attracting Capital via Narrative: Storytelling vs. Lying
35:07 - Toastmasters and Single-Topic Debates: Sharpening Communication
37:20 - Modern Credentials: Domain Expertise and Sourcing True Signals
44:23 - Government Sales 101: Enterprise Sales on Steroids
46:34 - Closing

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬3🔄4❤️8👀13102📊4

⚡ Powered by xgo.ing

honestly one of my favorite products rn and use them constantly especially while traveling. Love the...

andrew chen (@andrewchen) — Wed, 24 Jun 2026 06:04:51 GMT

honestly one of my favorite products rn and use them constantly especially while traveling. Love the new update! Congrats @davidsven and team!

David Woodland@davidsven

Today we’ve expanded our AI glasses portfolio with 26 more styles and colors.

Meta Glasses available today, starting at $299.

🔗 View Quoted Tweet

💬9🔄2❤️61👀22734📊11

⚡ Powered by xgo.ing

it’s said that to be a true science, then it must be testable and falsifiable. The scientific method...

andrew chen (@andrewchen) — Mon, 22 Jun 2026 00:39:13 GMT

it’s said that to be a true science, then it must be testable and falsifiable. The scientific method (hypothesis->experimentation->theory) can then be used to discover truth within these testable domains

There must be a deep connection to AI’s mastery of verifiable truths. That is, there’s an AI-native version of the Scientific Method where if a system is highly verifiable (AlphaGo, Erdos problems, etc) then AI will eventually know everything. And there are domains that might be in-between (it may turn out lots of science is like this). We’re use embodied AI/robotics to move the bar over time. Then there be some areas which are inherently non-verifiable and/or irreducibly complex, which is the domain of taste and human essence

Just as falsifiability determines the line between science and “not science” it feels like verifiability becomes the line between “AI can do it” and “needs a human”

💬15🔄1❤️43👀9962📊19

⚡ Powered by xgo.ing

Currently reading…

andrew chen (@andrewchen) — Sun, 14 Jun 2026 15:25:15 GMT

Currently reading…

💬13🔄0❤️35👀10226📊14

⚡ Powered by xgo.ing

Many of y’all know that the a16z Speedrun team puts on Tech Week in NY/BOS and SF/LA each year. It’s...

andrew chen (@andrewchen) — Mon, 08 Jun 2026 20:10:39 GMT

Many of y’all know that the a16z Speedrun team puts on Tech Week in NY/BOS and SF/LA each year. It’s where our startups do their GTM work, hiring, we meet founders, and bring together the ecosystem.

Just concluded the east coast with 50,000+ attendees and 2,000+ events

Incredible to see how much tech has grown in NYC, reflected in the numbers. Great start to our effort in Boston too

Next up, SF/LA in a few months.

Katia Ameri 🗽 #NYTechWeek@KatiaAmeri

That's wrap on Boston and New York Tech Weeks! We set new records across the board: 

🏆 New York Tech Week 
→ More than 5x growth since the first New York Tech Week launched four years ago 
→ 1,500+ events (largest Tech Week in history) 
→ ~50,000 attendees 
→ 400+ events in a single day (a Tech Week record)  

🏆 Boston Tech Week
→ 650+ events
→ 16,000+ attendees
→ An incredible turnout for our inaugural year  

Big shoutout to the team behind the success of these weeks!! @roseajohnson@ShrikalaKashyap@andrewchen@TracyMassaro_@filgran_@johnsonjanellea

See you in October for SF Tech Week + LA Tech Week! And keep up with us in the meantime @Techweek_ ✨

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬16🔄1❤️71👀13689📊16

⚡ Powered by xgo.ing

Lots of population decline/TFR discourse happening but we should factor in all the upcoming technolo...

andrew chen (@andrewchen) — Mon, 08 Jun 2026 17:03:23 GMT

Lots of population decline/TFR discourse happening but we should factor in all the upcoming technologies:
- longevity
- artificial wombs
- extending/rejuvenating fertility
- generating embryos 

Plus much tech to drive down the cost of families
- robotics
- automating housing construction
- AI teachers
- etc

Obv lots more that could both extend lifetimes, make it cheaper to have big families etc coming in the next few decades. Some will happen some may need more time. All feels inevitable though?

💬15🔄4❤️40👀17298📊20

⚡ Powered by xgo.ing

Short fantastic reminder of how spot on Kurzweil was - written in 1999, predicting AGI before 2050 s...

andrew chen (@andrewchen) — Sat, 06 Jun 2026 01:23:27 GMT

Short fantastic reminder of how spot on Kurzweil was - written in 1999, predicting AGI before 2050 simply extrapolating on Moore’s law versus neurons in the brain, before transformers and modern AI

Just a ~2 hour short listen for folks who want a refresher from over 20 years ago

💬20🔄4❤️74👀10113📊33

⚡ Powered by xgo.ing

Original quote: "When women love us, they forgive us everything, even our crimes. When they do not ...

andrew chen (@andrewchen) — Thu, 04 Jun 2026 18:44:24 GMT

Original quote:

"When women love us, they forgive us everything, even our crimes. When they do not love us, they forgive us nothing, not even our virtues." - Balzac

💬2🔄0❤️21👀4117📊5

⚡ Powered by xgo.ing

When users love us, they forgive us everything, even our bugs. When they do not love us, they forgiv...

andrew chen (@andrewchen) — Thu, 04 Jun 2026 18:44:10 GMT

When users love us, they forgive us everything, even our bugs. When they do not love us, they forgive us nothing, not even our free credits
-Balzac or someone like that

💬30🔄2❤️74👀6835📊32

⚡ Powered by xgo.ing

Historic moment happening via AI right now in the LA mayoral race - there’s been amazing use of AI v...

andrew chen (@andrewchen) — Mon, 01 Jun 2026 23:27:37 GMT

Historic moment happening via AI right now in the LA mayoral race - there’s been amazing use of AI video gen with viral new clips getting dropped every few days. It might literally let @spencerpratt win a major city election starting from behind

More important this is very much the preview of what the future looks like. At the moment, you might know that election budgets are insane. They’ll raise $50-100m for a big city/state race, then immediately turn around and spend 90% of it on TV advertising, mail etc. why? The audience is the boomers that vote, not nerds

But this will change and evolve. The voting population is getting more digitally savvy. AI videos generated by voters will dominate with ad spend meant to boost but not supplant the viral content. Daily viral AI video battleground. Gonna be wild.

Voter generated content not UGC :)

Your browser does not support the video tag.

🔗 View on Twitter

💬24🔄12❤️103👀9208📊32

⚡ Powered by xgo.ing

slide decks = adult picture books

andrew chen (@andrewchen) — Sun, 31 May 2026 22:15:33 GMT

slide decks = adult picture books

💬30🔄7❤️124👀11495📊38

⚡ Powered by xgo.ing

the team built a way for founders and investors to get intro'd at Tech Week! Use it May 26 (today) t...

andrew chen (@andrewchen) — Tue, 26 May 2026 15:47:34 GMT

the team built a way for founders and investors to get intro'd at Tech Week! Use it May 26 (today) to Jun 7 to meet connect

try it here: intro.tech-week.com

how it works:
- Founders sign up and say what they're building
- VCs hit the same link for investor access, then browse and reach out directly

why?
tens of thousands of people come to Tech Week every year to build, launch, hire, etc.

~30% of Tech Week attendees are founders. Many are building the next big thing. But since Tech Week is decentralized, there's been no central place for founders and VCs to find each other.

So we build this tool, which is only available for two weeks only during Boston + NY Tech Week (from May 26-June 7)

enjoy!

💬46🔄48❤️423👀837827📊136

⚡ Powered by xgo.ing

for founders in Boston later this month - we’re hosting a few events for new startups + lots of stuf...

andrew chen (@andrewchen) — Fri, 22 May 2026 03:05:38 GMT

for founders in Boston later this month - we’re hosting a few events for new startups + lots of stuff on the calendar too 👇🏼

Tech Week@Techweek_

if you're raising a seed round this year, 4 boston tech week events that will actually help:

→ The Investment Game with @emilybenn12 (@a16z). watch three founders pitch live and see exactly what an a16z partner cares about. you'll spot patterns in your own pitch in 30 minutes. may 28. partiful.com/e/ytxnudn6nO1v…

→ Founders & Funders VC Reverse Pitch. investors pitch you their thesis, check size, what they look for. previous editions produced term sheets. partiful.com/e/GCNJtSkNtbbN…

→ Two Lanterns 2LVC x OSS Investor Dinner. open source founders + investors in a dinner format. lower pressure than a pitch event, higher signal than a happy hour. may 26. partiful.com/e/N0GDRp8irDf7…

→ Finta. How Founders Actually Raise in 2026. The real playbook on what's working right now. less aspirational, more operational. may 27. partiful.com/e/zdC0kojFy1dJ…

🔗 View Quoted Tweet

💬9🔄4❤️56👀24838📊12

⚡ Powered by xgo.ing

(This was my reaction after watching the Google keynote)

andrew chen (@andrewchen) — Wed, 20 May 2026 04:31:09 GMT

(This was my reaction after watching the Google keynote)

💬0🔄0❤️12👀4574📊1

⚡ Powered by xgo.ing

is agentic coding in terminal vs in the desktop IDE like the new “emacs vs vim” debate?

andrew chen (@andrewchen) — Wed, 20 May 2026 04:29:40 GMT

is agentic coding in terminal vs in the desktop IDE like the new “emacs vs vim” debate?

💬56🔄3❤️85👀11279📊54

⚡ Powered by xgo.ing

dear lazy web who’s on the Google “Advanced Protection” program and whatcha think?

andrew chen (@andrewchen) — Wed, 20 May 2026 04:06:25 GMT

dear lazy web
who’s on the Google “Advanced Protection” program and whatcha think?

💬8🔄0❤️22👀8616📊9