andrew chen(@andrewchen)

p(China) is what we should be talking about, not p(doom) What is the probability that China - with ...

andrew chen (@andrewchen) — Mon, 19 Oct 2026 19:00:12 GMT

p(China) is what we should be talking about, not p(doom)

What is the probability that China - with its lead in robotics, manufacturing, construction capability, and second place on AI, GPUs, innovation - will become the primary beneficiary of the technological revolution the world is experiencing, over the next decades?

If your p(China) is high, what are you doing - and what should we be doing - to advance US capabilities in the coming decades so that we can lead, not follow?

IMHO p(doom) is the sideshow

💬21🔄3❤️61👀8992📊26

AI doomers: “the world is gonna end in 10 years” the same AI doomers: “I must buy a house in presid...

andrew chen (@andrewchen) — Mon, 19 Oct 2026 16:15:09 GMT

AI doomers:
“the world is gonna end in 10 years”

the same AI doomers:
“I must buy a house in presidio heights / noe / mill valley ASAP!!”

Reminds me of davos billionaire climate doomers who fly private and buy beachside

Rohin Dhar@rohindhar

Number of $5M+ single family home sales in San Francisco (twelve month rolling median)

💬20🔄4❤️130👀11818📊24

I totally understand this

andrew chen (@andrewchen) — Fri, 16 Oct 2026 15:15:58 GMT

I totally understand this

RTVisual@RT_Visual_on_X

Chinese bride gifts her partner a flagship GPU at wedding — he breaks down in tears

She handed him an ASUS ROG Astral RTX 5090 that costs around $12,000

Your browser does not support the video tag.

💬60🔄31❤️820👀118754📊116

Just wow. What a demo.

andrew chen (@andrewchen) — Fri, 16 Oct 2026 04:41:08 GMT

Just wow. What a demo.

💬13🔄3❤️86👀22557📊19

link: https://t.co/1FbrYUjJbq

andrew chen (@andrewchen) — Mon, 12 Oct 2026 19:37:22 GMT

link: speedrun.a16z.com/apply

💬8🔄0❤️19👀3038📊9

HERE’S THE SIGN YOU’VE BEEN WAITING FOR- for those of you with side projects, late night coding ses...

andrew chen (@andrewchen) — Mon, 12 Oct 2026 19:36:01 GMT

HERE’S THE SIGN YOU’VE BEEN WAITING FOR-

for those of you with side projects, late night coding sessions, the jam sessions with friends/coworkers. Maybe already know you’re a founder- you who follow all the greats, try all the latest AI tech, and have always known you are going to start something. Here’s your chance to build faster and bigger, learn a crazy amount in a short time, and really test yourself in the real world.

Bet on yourself. Build your startup with us.

SPEEDRUN 008 IS HERE

Your browser does not support the video tag.

💬49🔄16❤️247👀598339📊79

:)

andrew chen (@andrewchen) — Mon, 12 Oct 2026 17:01:38 GMT

:)

💬0🔄0❤️19👀3346📊2

SPEEDRUN 008 APPLICATIONS NOW OPEN we are now taking intros/recos for founders we should be meetin...

andrew chen (@andrewchen) — Mon, 12 Oct 2026 16:49:59 GMT

SPEEDRUN 008 APPLICATIONS NOW OPEN

we are now taking intros/recos for founders we should  be meeting for SR008

We have a new SF office where we’ll be hosting the  residency, and I will personally be working with founders  on growth/distribution/fundraising/more. Apply now for a  Jan 2027 start

And yes, we do invest up to $1M but we do lots of other things for you too :)

more info: sr.a16z.com

—
SPEEDRUN 008
We invest up to $1M in your new startup.

Applications for SR008, our next residency, are open now through November 1, 2026. SR008 runs February 1 to April 16, 2027, in San Francisco at our new offices in Jackson Square

We help founders with capital, customers, and talent. When you join speedrun, you get:

- Capital: Up to $1M, plus $10M+ in credits from 200+ partners through the speedrun Marketplace. When you’re ready to raise your next round, Demo Day puts you in front of a room full of investors.
- Customers: Introductions to executives and customers through SR Insiders, our network of 150+ C-suite founding members and 1,000+ executives and operators. You’ll also get AI Faire and Tech Week to meet buyers in person, Launch Lab to get your product in front of people, and Brand Lab to shape how you show up from day one.
- Talent: Access to the Talent Network of 171,000+ people, which our team is always growing, plus hiring and HR support as your team scales.

You’ll also be building alongside a community of 1,000+ speedrun founders who support each other with customer introductions and product feedback.

💬74🔄55❤️722👀1525419📊212

FINAL day to apply for the Speedrun Alpha fellowship https://t.co/aiNkBstBKd For students / dropou...

andrew chen (@andrewchen) — Mon, 12 Oct 2026 01:42:36 GMT

FINAL day to apply for the Speedrun Alpha fellowship

speedrun.a16z.com/alpha

For students / dropouts / recent grads: we want to invest up to $250k. Come work out of our SF offices and explore ideas, find a cofounder, etc.

More details:

Start building. We back you pre-entity, pre-idea, pre-team. Apply as an individual. If you already know who you want to build with, great. Have each person apply separately and note who you’re building with.

- Up to $250K investment
- Automatic final interview for speedrun, up to $1M investment
- $1M in credits across AWS, GPC, Azure, Open AI, Anthropic, Cursor and many more.
- Coworking space at the speedrun office throughout the fellowship
- Office hours and hosted sessions with speedrun partners
- 8-week, in-person experience (January 11 – March 5)

💬16🔄5❤️90👀197842📊33

important points on the future of LA vs SF The sad part of LA is that the movie business jobs are p...

andrew chen (@andrewchen) — Sun, 11 Oct 2026 20:16:24 GMT

important points on the future of LA vs SF

The sad part of LA is that the movie business jobs are predominantly blue collar workers that build and operate the industry, not just celebrities. And it’s sad that those jobs are going away

Pirate Wires@PirateWires

San Francisco is cleaning up its act while LA doubles down on its worst policies — and as a result, tech is staying in SF, but Hollywood is leaving LA.

Your browser does not support the video tag.

💬9🔄3❤️65👀16702📊13

wow fleet week. traffic in sf so messed up today that my Waymo was blocked for 15 min and had to esc...

andrew chen (@andrewchen) — Sun, 11 Oct 2026 19:31:25 GMT

wow fleet week. traffic in sf so messed up today that my Waymo was blocked for 15 min and had to escalate to humans in the loop to reroute. Made a pretty hilarious left turn and big loop, finally on my way back. 20 min ride turning into a 50 min detour

💬20🔄0❤️93👀16605📊25

Hosting our first hackathon at the Speedrun offices in Jackson square to close out SF Tech Week! Se...

andrew chen (@andrewchen) — Sun, 11 Oct 2026 19:28:20 GMT

Hosting our first hackathon at the Speedrun offices in Jackson square to close out SF Tech Week!

See y’all in LA next!!

speedrun@speedrun

hour 1 of SpeedHacks 💻  sharing more throughout the day...

💬14🔄6❤️94👀9843📊17

the new waymo ojai vehicle 👌🏼👌🏼👌🏼👌🏼 it’s so nice. Finally got one today in SF

andrew chen (@andrewchen) — Sun, 11 Oct 2026 05:27:12 GMT

the new waymo ojai vehicle 👌🏼👌🏼👌🏼👌🏼

it’s so nice. Finally got one today in SF

💬35🔄0❤️129👀31097📊40

btw Tech Week in SF has grown 10x since we started running it in 2023 Inevitable that this will eve...

andrew chen (@andrewchen) — Thu, 08 Oct 2026 23:36:03 GMT

btw Tech Week in SF has grown 10x since we started running it in 2023

Inevitable that this will eventually be the largest tech conference in the world

💬34🔄9❤️170👀17079📊47

some fun stats from SF Tech Week - the largest ever - with thousands of events and 50k+ attendees. W...

andrew chen (@andrewchen) — Thu, 08 Oct 2026 21:13:36 GMT

some fun stats from SF Tech Week - the largest ever - with thousands of events and 50k+ attendees. We've more than doubled since we started hosting in SF

Amazing to see FiDi emerge as a force! Used to be that tech was just in SOMA. Plus the Mission+Dogpatch as well

💬18🔄14❤️142👀17735📊38

👀👀👀

andrew chen (@andrewchen) — Thu, 08 Oct 2026 16:40:58 GMT

👀👀👀

💬11🔄3❤️178👀57822📊39

fun to use town/muse/grokbot after months of being a DAU on hermes/openclaw Reminds me of all the a...

andrew chen (@andrewchen) — Wed, 07 Oct 2026 03:13:47 GMT

fun to use town/muse/grokbot after months of being a DAU on hermes/openclaw

Reminds me of all the amazing fun and insanity it is to admin your own linux box versus using something like iOS

Not switching for now! Too much fun nerding out

💬42🔄0❤️118👀15702📊47

Link to more writing: https://t.co/2JRuUM1TDf

andrew chen (@andrewchen) — Tue, 06 Oct 2026 15:09:43 GMT

Link to more writing:
andrewchen.substack.com

💬0🔄0❤️5👀2464📊1

DO AGENTS HAVE NETWORK EFFECTS? We are seeing a new wave of general purpose consumer-friendly a...

andrew chen (@andrewchen) — Tue, 06 Oct 2026 05:18:59 GMT

DO  AGENTS  HAVE  NETWORK  EFFECTS?

We are seeing a new wave of general purpose consumer-friendly agents, and it’s causing many thousands of AI startups to re-evaluate their place in the market. Will these new horizontal players take over completely? Will it be winner take all? Can products that focus vertical use cases survive? The new wave of products - Muse, Town, Instinct, Grokbot, etc - come from taking the power and magic of Openclaw/Hermes but wrapping it in polished, secure, consumer-friendly experience. This is only the beginning, and soon every major tech company will throw their hat in the ring and thousands of startups that have been working on vertical agents in X (travel, shopping, family, work, SMB, etc etc) have big existential questions to answer.

The simple case against winner-take-all dynamics: These agents are tools, not networks. They inherently provide a solo experience, and although packaged as chat experiences, they don’t have network effects the way that Whatsapp or phone networks do. They’re more like email clients - I can use Superhuman and someone else uses Outlook and someone else is on Gmail, but in the end we just send email to each other. In that metaphoi, if a Muse agent encounters a Instinct agent, they will just figure out how to talk to each other by building APIs, or using email/messaging/etc., so that there’s no advantage to everyone running the same thing. Further, you can swap out the underlying LLM - as we saw in the Claude vs Clawdbot saga - things will still work. The memory and underlying context are just text files, and AI are great at porting things from one system to another. So where’s the moat?

Just because they are tools today doesn’t mean they can’t evolve into networks. As I’ve written about in the past, instead of discussing “network effects” as a vague/amorphous concept, let’s instead overlay it against core KPIs each product has to build against:

- Acquisition network effects: More users** = lower CAC, higher virality, more signups
- Engagement effects: More users = higher retention, deeper usage, more frequency
- Monetization effects: More users = higher ARPU, higher conversion, more share of wallet
(**PS. and by “more users” of course I actually mean, of course, higher network density with more interconnection within/across networks. Not scale effects)

So looking at these sub-categories of network effects, you might ask: What are the features that you’d build to actually create network effects in agents? How does this category become winner-take-all?

Acquisition:
In the pre-AI world, network effects in customer acquisition were driven by users taking actions that then bring in even more users - whether that’s sending an invite, sharing a piece of content, or getting mentioned in a comment. Those users would then join the network, repeat the same actions, creating more viral loops that would throw off more users. In the post-AI world, both users and agents can initiate these loops. An agent might suggest, “hey, do you want me to help organize a New Years celebration with family” and then might pull together a group chat with your siblings, in-laws, and parents. It might build a microsite for the event and ask people to sign up and RSVP, creating an opportunity to generate a signup. Or for work, you might finish a slide deck and your agent asks if you want to pull together a meeting to present it to your team. When it does that, it might create a prep doc, take notes, and host it all on a microsite that engaged others.

These types of agentic viral loops need a few things to work: They need to be able to plug into your contacts, and to have enough context to know when to loop which people into which projects/events/activities. They need access to communication channels like email/messaging/otherwise, specifically where they can reach potential users, as this serves as the viral substrate for propagation. The magic will be how to get all of these things without being creepy, and to create enough trust to reach out to friends/colleagues/etc on your behalf, without being too push. No one wants an agent that says, “hey Andrew would like to get lunch next week” followed by “and sign up for this agent to agree!” - it has to be better and more subtle than that.

Many viral loops in the pre-AI era, particularly in the workplace, were built on a content creation loop - you use Google Slides to make a thing, you share a link, and then people would view it. Eventually they might make their own. Many products like Figma, spreadsheets, Notion, Canvas, are all built on this loop, and so are YouTube, Instagram, etc. In the agentic world, you might start by talking to your agent about interior decoration ideas, and it might ultimately build you a Pinterest-like board and make it easily shareable with others. It might build you a spreadsheet mini-app with a budget, a schedule, etc., and again, make it easily shareable. It might volunteer to do this inside a group chat with your friends, and if they want to view it, they should sign up and get their own agent too.

If the content creation loop works for agentic virality, I think it means there’s a big incentive for every agent to volunteer to make visually-appealing, shareable content, do this often. AI is so good at codegen that creating a one-off disposable artifact will probably be more useful than not. And there’s a big incentive to drive virality by building/owning the content yourself as the agent. In the interior decoration example, it would be more viral to create a standalone version hosted by the agent, rather than building an actual Pinterest board, or to build a self-contained mini-spreadsheet rather than creating a GSheet. A properly tuned agentic viral loop will do the former because it helps spread the agent, rather than helping spread Pinterest/GSheets.

This might all sound like horrific spamminess to you, but I think it might be done tastefully. If an agent is high-retention and high-usage, you’ll have many shots on goal to gradually invite your friends onto the same platform as you. The first shot might come from an invite loop that says your colleague/friend wants to set up a place to share photos/coordinate calendars/etc., and even if you say no, over time, your friend might share very useful microsites for projects or events or research that’s relevant, and you might decide that you want to sign up to leave a comment, but then eventually you might try the other functionality as well.

Engagement:
It’s easy to imagine that in a world where everyone by default has super intelligent agents that product engagement in agents would go up. The simple argument is that the ubiquity of agents would lead to more successful task completion, making agents more useful and applicable to more problems. Routine tasks like scheduling something or summarizing/sharing notes from a meeting, would happen instantly if it ran through agents and you removed humans completely. As agents become more successful over time, not only with your own tasks but in coordinating larger and more complex multi-user decisions, you’d end up using them more.

However, this does not answer the question of network effects. The question there is, does your agent get better when other people are all using the SAME agent? As I said earlier, the argument for NO is if each type of agent runs on a frontier AI model, can talk to other agents (all different types) in real-time via email/messaging/APIs/connectors/etc. Then there’s not much leverage. If you’re just organizing a meeting or a dinner party or meeting the agents just ask each other in real-time - and even if they are heterogeneous - they can use email and calendars and figure it out.

But some tasks do benefit from everyone running the same agent, because you get the benefits of centralization: Trust, speed, identity, discovery, etc. Some examples:

-  “Who are 10 people in San Francisco that I should meet?” - agents can’t just ask each other to find an optimal solution
-  “Find me a VP of Engineering who isn’t publicly looking for a new job, but would leave for a fast-growing startup that pays X in stock” - there’s a mix of private and public info here, and an agent of a passive jobseeker might lie if the asking agent is not trustworthy
-  “Who’s a trustworthy seller of X that I should buy from?” - reputation might be gamed, and it might be better to have a global source of truth
-  “Help me pick out the best dinner in the next 5 minutes that my friends and all their +1s would like, incorporating allergies/schedules/locations” - real-time solution incorporating a bunch of shared context
-  … and even more so, if you imagine that there may be agents that lie, exaggerate, or otherwise to benefit their humans

For the above examples, you could imagine Muse/Town/Instinct/whatever building a shared network+context layer that understands the network, its relation to public/private data, acts as a system of record for identity/trust, does multi-sided marketplace matching, etc. Then it would allow only its agents access to this platform, and agents outside of the network wouldn’t have access. If you’re part of this network, and are contributing data, your agent gets more useful over time. The network gets more powerful over time.

The counterpoint to this counterpoint is that perhaps new startups will fill the void as a series of agent-facing point solutions, making it possible for people running agents of all times to interface with each other. Perhaps Muse creates a proprietary ecommerce marketplace that only Muse agents can use, but perhaps there will just be a next-gen eBay for any agent to sell to any other agent, and this service maintains central reputation/identity at that layer. In other words, perhaps the centralization will still happen, but just at a layer above agents. Most likely both will happen - every large tech company adds chat, commerce, search, etc, but also there are pure play versions of each of these. But this feels like more fragmentation than a true winner-take-all.

Some parting thoughts:
My argument in this writeup is simple: Agents don’t inherently have network effects. If all we build are super-intelligent assistants that can use tools and talk to each other, then the category might look surprisingly fragmented. Your agent can talk to my agent, just as Gmail can send email to Outlook, and there’s no particular reason we need to use the same one. Better models, UX, integrations, memory, and distribution might create very large companies, but those are not network effects.

But the next gen of agents have a big opportunity design network effects in their interactions. On acquisition, they can turn the things they create—documents, events, websites, research, group chats, plans—into viral objects that naturally pull other people in. And on engagement, they can create shared networks of identity, reputation, relationships, private context, and intent that make the agent increasingly useful as more of the people around you join.

There’s many ways for these network effects to instantiate. If we’re lucky, things will look more like the open web - mostly decentralized, with bits of centralization to make certain tasks smoother. Identity could become an open protocol. Reputation could live in marketplaces. Commerce could happen through an agent-native eBay. Professional discovery could happen through an agent-native LinkedIn. Agents could remain interchangeable clients sitting on top of a constellation of networks and marketplaces, much as browsers and email clients do today. In that world, enormous network effects emerge around agents without producing a winner-take-all market for the agents themselves.

My guess is that we’ll see a fight over exactly this boundary. Every major horizontal agent will try to pull valuable network functionality inside its walls: host the artifact instead of creating a Google Doc, execute the transaction instead of sending you to Amazon, make the introduction instead of searching LinkedIn, create the event instead of using Partiful. Every time it does this successfully, it converts somebody else’s network effect into its own. Meanwhile, every incumbent network and a new generation of startups will have the opposite incentive: make their networks universally accessible to every agent so that no horizontal agent can recreate and capture them.

This is why I don’t think the interesting question is simply, “Will agents have network effects?” Of course some will. The much more consequential question is: Where will the network effects live? If they live primarily in open protocols and specialized networks, we may end up with thousands of agents competing on intelligence, personality, UX, specialization, and price, all interoperating with the same underlying ecosystem. Vertical agents can thrive in this world. But if a few horizontal agents successfully accumulate identity, relationships, private context, reputation, latent intent, and distribution - and keep those assets proprietary - then switching agents starts to mean leaving your network behind. Next couple years will be very fun to see how this plays out.

💬163🔄54❤️638👀502603📊368

SF Tech Week kicks off today!! We’ll be hosting a bunch of events over the next few days, at our new...

andrew chen (@andrewchen) — Mon, 05 Oct 2026 15:56:47 GMT

SF Tech Week kicks off today!! We’ll be hosting a bunch of events over the next few days, at our new Jackson Square office and for our massive Demo Day with 1000+ investors.

Also sharing a couple fun stats/factoids 👇🏼

-
SF TECH WEEK 2026
1,700+ events on the calendar, more than double (!!) in the past two years

1,266 hosts from 1,058 companies, including OpenAI, Anthropic, SpaceXAI, Stripe, Google, Vercel, Databricks, ŌURA, Mistral, Perplexity, Cerebras, Lovable, Ramp, Brex, Mercury, AWS, Adobe, Reddit and more

100+ events hosted by portfolio companies of Speedrun, a16z

Fastest-growing themes:
- Hardware: tripled, the strongest and most consistent growth of any theme
- Cybersecurity: more than tripled from a small base, largely driven by AI security as companies deploy agents
- Infrastructure: more than doubled, as compute and the AI stack draw more builders
- Deep tech: doubled, with more science- and research-led startups on the calendar

💬26🔄20❤️159👀18612📊48

Calendar lives here: https://t.co/cnf6SekySS

andrew chen (@andrewchen) — Fri, 02 Oct 2026 00:50:39 GMT

Calendar lives here:
tech-week.com

💬3🔄2❤️16👀3864📊6

who’s in SF next week? We’re hosting, co-hosting, and organizing a bunch of events with great speake...

andrew chen (@andrewchen) — Fri, 02 Oct 2026 00:49:40 GMT

who’s in SF next week? We’re hosting, co-hosting, and organizing a bunch of events with great speakers including at our new Jackson Square office.

Full SF+LA tech week calendar now online!

💬42🔄15❤️272👀23390📊76

welcome to speedrun @abliteration_ai team and @founderengineer! "AI with guardrails that you contro...

andrew chen (@andrewchen) — Thu, 01 Oct 2026 16:30:25 GMT

welcome to speedrun @abliteration_ai team and @founderengineer!

"AI with guardrails that you control" - love it.

💬10🔄4❤️59👀15116📊15

when the personal computer won, it made $20k+ workstations in 1980s a distant memory - SGI, Next, Su...

andrew chen (@andrewchen) — Wed, 30 Sep 2026 18:57:07 GMT

when the personal computer won, it made $20k+ workstations in 1980s a distant memory - SGI, Next, Sun, etc. I remember thinking, huh why would anyone ever buy a computer that expensive!??

yet now you buy a PC and add a single RTX PRO 6000... or a Mac Studio. And prices are only going up...

💬44🔄3❤️138👀22465📊52

yup. This is where open weight AI needs to go Customer controlled guardrails >>> black box...

andrew chen (@andrewchen) — Wed, 30 Sep 2026 01:20:26 GMT

yup. This is where open weight AI needs to go

Customer controlled guardrails >>> black box rules

💬18🔄3❤️42👀14807📊20

a fellowship with $250k for recent grads/students - deadline next week link: https://t.co/w7mfnRgJ...

andrew chen (@andrewchen) — Tue, 29 Sep 2026 15:18:22 GMT

a fellowship with $250k for recent grads/students - deadline next week 

link: alpha.a16z.com

more info 👇

--
Applications are closing soon for a16z speedrun alpha, our new fellowship for technical students and recent grads who want to start building now

- Up to $250K investment
- $1M+ in credits across AWS, GPC, Azure, Open AI, Anthropic, Cursor and many more.
- Automatic final interview for speedrun, up to $1M investment
- Office hours and hosted sessions with speedrun partners
- Coworking space at the speedrun office throughout the fellowship
- 8-week, in-person experience (January 11 – March 5)

for folks that would rather get experience before starting their company, we also have a Talent Track that fast-tracks you toward roles at breakout speedrun startups

just over a week left to apply:alpha.a16z.comj

💬19🔄26❤️290👀2096108📊101

yes, but we need more... a lot more

andrew chen (@andrewchen) — Mon, 28 Sep 2026 02:07:08 GMT

yes, but we need more... a lot more

Patrick Collison@patrickc

Some data we recently assembled on entrepreneurship/compute in Europe: eudata.vercel.app.

We hope that one of the useful roles that Stripe can play is in collecting and publishing empirical data pertaining to entrepreneurship and industry in Europe. There's growing appetite to get Europe on a better footing, and cross-sectional comparisons can often shine light on where opportunities lie. If you're interested in this kind of thing, we publish more at stripeeconomics.substack.com.

💬29🔄12❤️159👀44621📊38

Every graph in this article is amazing. https://t.co/KIk1rmze26

andrew chen (@andrewchen) — Thu, 24 Sep 2026 03:09:37 GMT

Every graph in this article is amazing.

wsj.com/economy/the-ai…

💬16🔄22❤️200👀36197📊53

Epic

andrew chen (@andrewchen) — Thu, 24 Sep 2026 03:08:05 GMT

Epic

💬136🔄398❤️2186👀5996568📊564

Love that there’s a bunch of new startups focusing on local AI / sovereign AI / GPU hoarding 😂 spea...

andrew chen (@andrewchen) — Thu, 24 Sep 2026 00:06:21 GMT

Love that there’s a bunch of new startups focusing on local AI / sovereign AI / GPU hoarding 😂 speaking my language!

Congrats to @zainmfj and the @ghostai team

Zain Javaid@zainmfj

x.com/i/article/2102…

💬18🔄7❤️150👀26506📊35

by now we’re used to: - frontier model launches - 6-12 months later, the chinese open weight models ...

andrew chen (@andrewchen) — Mon, 21 Sep 2026 03:56:34 GMT

by now we’re used to:
- frontier model launches
- 6-12 months later, the chinese open weight models close the gap

this week we’re seeing:
- jev launches
- a few days later, swarms of US AI developers ship their own open weight models :)

tracking 2-3 early ones so far, but looks like some great open weight alternatives coming within the week

I’m sure there will be some great chinese open weight alts for jev coming also

💬76🔄14❤️303👀43370📊105

curious if Jev-like models might accelerate us towards on-device AI native functionality on mobile d...

andrew chen (@andrewchen) — Sun, 20 Sep 2026 22:29:07 GMT

curious if Jev-like models might accelerate us towards on-device AI native functionality on mobile devices

strong LLMs are a long way from running on phones:
- slow memory bandwidth
- only highly quantized MoE models will fit
- lots of issues with power/heat/etc

Folks are building NPUs for the next gen of mobile but targeting pretty modestly sized LLMs

There’s lots of mobile UX that would benefit from fast/cheap AI decision models - from notifications, typing/texting, to in-app experiences like inboxes/calendars/etc

I always figured the other solution would be a model-on-a-chip that would encode an older-yet-useful model into the actual phone hardware itself, but maybe this will move faster?

💬69🔄12❤️199👀21060📊88

I prev wrote about the battle of inference cost vs ARPU here - it might end up sooner than expected!

andrew chen (@andrewchen) — Sun, 20 Sep 2026 03:43:28 GMT

I prev wrote about the battle of inference cost vs ARPU here - it might end up sooner than expected!

andrew chen@andrewchen

for AI-native consumer apps to be truly ubiquitous we need:

ARPU > Average Inference Cost Per User.

How far away are we from that? Ideally AI native apps can hit APIs on every screen yet also pay for just by throwing some ads on it

But I think we’re a while away:
- unfort we’re probably >10x off right now. Monthly ARPU is $2-5? Token costs for an AI heavy app might be $20-50 of cost 
- much of consumer is global. Even if we hit the US/EU it’ll be a while before we can serve the broad base 
- more importantly every time AI improves, consumers demand more. No one wants to talk to a last gen AI character. If video gets good they’ll want videos everywhere 
- same issue for LLMs. We started with short replies back. Now we want agents who can do entire tasks for us
- it might be we need major innovations in small models or new mobile hardware so that we have free local inference. But that will still not be as good as SOTA cloud LLMs obv, but might solve some use cases 

No wonder so many products focus on productivity, and on prosumers who can pay $100s or $1000s on work related tasks. This is where you can have huge ARPUs and benefit from being SOTA. And it seems as though there’s no limit for tokens… so why do the low end?

💬3🔄1❤️27👀9830📊6

Jev is going to change the prosumer/consumer AI landscape by unlocking a specific thing: Ad-support...

andrew chen (@andrewchen) — Sun, 20 Sep 2026 03:42:19 GMT

Jev is going to change the prosumer/consumer AI landscape by unlocking a specific thing:

Ad-supported + free AI native apps

this is going to usher in a generation of new AI-native marketplaces, social networks, photo apps, messaging, calendars, email, collab tools, and much more.

Why? Previously, if you wanted to have an AI-native app with multiple/fast LLM calls on every screen, the math just didn’t work - inference costs just couldn’t be paid back from throwing a few video ads or affiliate links. So there were really just two solutions:

1) be a massive company and subsidize AI costs
2) charge a subscription fee to cover inference
(or both)

So what happens when you take a different approach that’s looks to be, initially, >400x cheaper than a more general LLM? I think it means that app developers are going to make their apps smarter and different in interesting ways. It won’t be as generically powerful as LLMs, but instead folks will create highly differentiated point solutions. You may not have a generically powerful virtual admin to schedule all your meetings, but you may have a free/ad-powered assistant that can comb through your inbox, filter for only important emails, figure out important dates, prioritize contacts, and do other interesting 80/20 work. Same for generically powerful virtual companions, or app builders, etc - it may be that we can hybridize the LLM and future Jev-like models that adds “wow” features at 1/1000ths of the cost, to make the economics work.

If folks hear of interesting new apps/interactions/etc built on Jev, I’d love to hear it!

💬135🔄81❤️1124👀102722📊439

Sometimes it’s the small victories that make your day

andrew chen (@andrewchen) — Wed, 16 Sep 2026 01:42:30 GMT

Sometimes it’s the small victories that make your day

💬34🔄0❤️176👀16763📊42

ok great now I have 3 messy inboxes I have to check: - stuff my agents have done on scheduled loop ...

andrew chen (@andrewchen) — Tue, 15 Sep 2026 02:12:04 GMT

ok great now I have 3 messy inboxes I have to check:

- stuff my agents have done on scheduled loop that I need to review

- bookmarked slack messages off Google Docs and slides to review

- my actual email inbox

lol

💬115🔄3❤️202👀27351📊117

current homelab setup for local AI experimentation: - hermes box hosted on a Framework Desktop Main...

andrew chen (@andrewchen) — Sun, 13 Sep 2026 21:21:03 GMT

current homelab setup for local AI experimentation:

- hermes box hosted on a Framework Desktop Mainboard AI Max+ 395
- 5090 eGPU running Qwen 3.8 27B for fast tok/s LLM use - sometimes 150+ tok/s
- 2x DGX Spark: running Deepseek v4 Flash 0731 - better but slower model
- pi 5 for monitoring
- Mac mini as a dev box - use Herdr and ohmypi/codex/claude depending on the use case
- housed in a 10" DeskPi mini rack (mostly)

Hermes is defaulted to local AI but with a homegrown routing plugin hitting a small low TTFT model (Arch-Router) to decide whether to go local or upgrade to cloud/frontier. Trying to get to 100% local over time, but right now probably more like 60-70%

The Sparks are for batch processing background runs (all the cron jobs, longer dev builds, etc)

do you need all of this? Absolutely not lol. I started with the mac mini and couldn't help myself but to add over time!

Also I regret getting the eGPU so I wouldn't recommend that to anyone

💬85🔄16❤️301👀35679📊147

https://t.co/C6l9JixsOk

andrew chen (@andrewchen) — Thu, 10 Sep 2026 18:06:33 GMT

tech-week.com/calendar

💬2🔄0❤️16👀3919📊6

hey all- for folks who are in town (SF) Oct 5-11, we're hosting a ton of events, including at our ne...

andrew chen (@andrewchen) — Thu, 10 Sep 2026 18:06:16 GMT

hey all- for folks who are in town (SF) Oct 5-11, we're hosting a ton of events, including at our new offices in Jackson Square.

details below.

--
the SF/LA Tech Week calendar is LIVE 

2,300+ events across the two weeks. more than a 50% increase from last year.

some other cool stats:
- 1,580 events in SF, 762 in LA
- 14 tracks - fundraising, deep tech, AI agents, infra, devtools, hackathons, media, fintech and more
- 883 events on fundraising & investing. more than all of LA Tech Week in 2024.
- 80+ a16z and portcos hosting: deel, stripe, elevenlabs, pinecone, databricks, carta, fal, exa
- spacex, erebor, anthropic, google, aws, ibm, fenwick, hsbc, fireworks, vercel, pwc, fin, adobe, atlassian, cloudflare, hyperagent all showing up

congrats @roseajohnson@TracyMassaro_@KatiaAmeri@filgran_@johnsonjanellea and the whole Tech Week team for the awesome launch

💬46🔄18❤️323👀323342📊95

What % of the content we consume in 10 years will be AI generated?

andrew chen (@andrewchen) — Wed, 09 Sep 2026 02:38:50 GMT

What % of the content we consume in 10 years will be AI generated?

💬56🔄5❤️31👀17813📊51

the incoming DM just flipped to coming from @ turtlesoupy but @ EnforcementX is still up.. wtf..

andrew chen (@andrewchen) — Tue, 08 Sep 2026 18:29:13 GMT

the incoming DM just flipped to coming from @ turtlesoupy but @ EnforcementX is still up.. wtf..

💬1🔄0❤️2👀1938📊1

Pretty bad impersonation from @ EnforcementX Please help me report them for impersonation of @Suppo...

andrew chen (@andrewchen) — Tue, 08 Sep 2026 17:45:09 GMT

Pretty bad impersonation from @ EnforcementX

Please help me report them for impersonation of @Support

💬6🔄4❤️38👀15001📊10

you want to spend zero time on the left. This is where AI wins - automating drudgery/workflows/costs...

andrew chen (@andrewchen) — Tue, 08 Sep 2026 05:51:35 GMT

you want to spend zero time on the left. This is where AI wins - automating drudgery/workflows/costs

we want to spend our time on the right. More of it is about connection/entertainment/humanity - less verifiable, more novelty-seeking

💬76🔄40❤️462👀41512📊189

at work, makes sense you have agents effectively represent your CFO, CMO, CRO, CPO, etc. -- then sup...

andrew chen (@andrewchen) — Fri, 04 Sep 2026 04:16:22 GMT

at work, makes sense you have agents effectively represent your CFO, CMO, CRO, CPO, etc. -- then supervised by a COO agent

What's the equiv for your personal?
- family
- personal finance
- health
- leisure/travel
... what else?

💬54🔄7❤️105👀19021📊68

i'd pay 2x for an iPhone where siri/dication/other AI features actually were good

andrew chen (@andrewchen) — Fri, 04 Sep 2026 00:24:41 GMT

i'd pay 2x for an iPhone where siri/dication/other AI features actually were good

💬95🔄4❤️350👀29991📊110

GTA7 is gonna be sick

andrew chen (@andrewchen) — Thu, 03 Sep 2026 22:16:05 GMT

GTA7 is gonna be sick

Matt Shumer@mattshumer_

GPT-6 Astra built this Manhattan world in Unreal Engine over the course of a week.

It was literally able to go street by street to make each one perfect.

Your browser does not support the video tag.

💬35🔄24❤️1669👀228437📊156

Pretty clear we're about to get a VR comeback, no?

andrew chen (@andrewchen) — Thu, 03 Sep 2026 20:41:39 GMT

Pretty clear we're about to get a VR comeback, no?

Thomas Ricouard@Dimillian

Astra is very good at 3D modeling, and I can't wait for all of you to experience it, for now here is a little walkthrough on how I built the demo house for our launch blog post. From a Blender scene to a Unreal Engine 5 walkable experience. 🧵

Your browser does not support the video tag.

💬145🔄66❤️1487👀156342📊251

2012: just add chat/DMs to your mobile app! 2026: just add notetaking to your AI app!

andrew chen (@andrewchen) — Thu, 03 Sep 2026 19:46:17 GMT

2012: just add chat/DMs to your mobile app!
2026: just add notetaking to your AI app!

💬15🔄1❤️36👀14783📊17