Pinned Tweet
Pass App is changing the game with AI agents.
First wallet that lets you say what you want and an agent makes it happen.
Huge unlock for newbies who don’t understand how to use crypto, and huge time saver for degens.
And this is just the beginning 🚀
🚀 Introducing the first intelligent wallet.
At Pass App, we’re redefining what a crypto wallet can do. Powered by AI agents, you can swap, bridge, buy, and send using simple voice or text commands.
No toggling chains. No hunting contract addresses. No clunky UX. Just say what you want and your agent handles the rest. You stay in control.
Read the full article 🔗👇
These are so good 👏
In light of Ami Vora becoming @Anthropic's Head of Product, here are my top 10 takeaways our recent interview.
📌 Watch our full interview here: youtu.be/Vp2BsJNVFqQ
We’re giving away lots of prizes this month. Will be a fun few weeks 😄💰🎁
🎁 DeFi December starts NOW
From now until Dec 24, we’re giving away over $3,000 in prizes — including a $1,000 grand prize.
✅ New prizes every week
✅ Every challenge = 1 entry to win the big bag
To kick things off? A quick win:
💜 Follow @passappHQ
🔁 RT this post
🎉 4 winners get $50 USDC each to their Pass App wallet
⏰ Giveaway ends in 48 hours
Tag a degen who deserves a little onchain holiday cheer 👇
1
128
Josh Ben-David retweeted
Have Pass App installed? Congrats, you’re already on the nice list.
Something big (and onchain) is coming to light up your holiday season 🎄✨
🗓️ 24 days of giveaways
💰 Over $3,000 in USDC up for grabs
🎁 $1,000 USDC grand prize
It all kicks off December 1.
Follow us so you don’t miss your shot to stuff your bags 👀
Josh Ben-David retweeted
TBH this is the easiest bear market I've ever seen.
Seems like most of you have forgotten what 2022 was like. Luna collapsing, then 3AC, then FTX, then Genesis, BlockFi, Axie, NFTs--pretty much everything felt like a house of cards.
And then after all that stuff collapsed, the BANKS collapsed, stablecoins depegged, Gary Gensler and the president tried to destroy almost every company in the industry--I really wasn't sure how much of this stuff was going to survive.
Compared to that? This is breezy.
Prices have gone down, yeah whatever. Fundamentals are great. Crypto is working.
So relax, get something to eat. Lock in. We'll be fine.
Since the second half of last year, a lot of my friends(OGs in crypto) left crypto, as they started to doubt their original belief in crypto/blockchain/Web3 like do projects really create value to the world, what’s profit market fit, only speculations like meme/perp/prediction projects survive in this cycle, do web3 projects like infra/social really have values?
Right now, it’s almost the toughest mode for startups, vcs, traders and users. Narrative play is not working, and hot topics shift every week. Organic liquidity is drained to lower and lower on cex and dex. Even without listing fees, few ppl want to buy new/old tokens. Market manipulators create pump and dump with leveraged perp to operate small cap or old coins every day.
We can always say defi or trading is the only pmf for crypto/blockchain. But OGs and builders like @VitalikButerin @pmarca always paint a bigger vision of world computer/web3 to not replace but at least build a better internet or crypto eat the world.
With llms like @OpenAI @claudeai, not just crypto but all industries are facing a 100-1000 to 1 problem: one experienced person can do 100-1000 ppl’s job with llm’s help. New tools and products are releasing too fast and will be faster for the next 1-2 years. Not just in crypto but in web2, tech and engineering start to become admin work(assigning tasks to llms to complete).
No matter for vcs and startups, keep pivoting and shipping is the only way out, until can dominate the particular sector. Staying relevant is the easy to say but hard and have to achieve to survive in this new meta.
My Ph.D. advisor always says persistence pays off. Hope we can win this uphill battle. To friends, to ogs, to builders, to dreamers, to crypto.
Based on my chats for the past few months with @VitalikButerin @tkstanczak @hosseeb @sreeramkannan @tarunchitra @ASvanevik @StaniKulechov @Lomashuk @lukejdpearson @13yearoldvc @YeruiZhang @jocyiosg @artofkot @no89thkey @xingpt @broleon
Josh Ben-David retweeted
Replying to @PassAppHQ @HyperliquidX
Whether you’re new to perps or a seasoned degen, Pass App gives you the tools to trade with confidence, minus the chaos.
Read more in our blog: pass.app/blog/perps-trading-…
If you're sitting comfy in stables waiting for the bounce, this is @PassAppHQ right now 😯
NFA/DYOR
Josh Ben-David retweeted
BREAKING: Grok just claimed the #1 spot on the Roo Code Daily and Monthly leaderboards! 🏆
Great read. Many opportunities ahead. Hope humanity doesn’t lose its humanity during this transition.
THIS IS WHAT'S KEEPING ME UP AT NIGHT:
1. AI will kill the concept of a 9–5 for millions. MANY get laid off, become freelancers, shift to portfolios of agent-assisted work.
2. livestreaming explodes 100×. it becomes the only way to prove you are real and not AI. Twitch will look like one of the greatest acquisitions of all time.
3. the creator economy is graduating into the founder economy. audiences are mobilizing into companies, funds, and franchises. MrBeast was just the prototype!
4. we’re entering the app recombination era. the biggest startups of 2026 will be built by remixing three or four existing AI tools into new vertical workflows.
5. agents will start talking to other agents, and you won’t be in the loop. every “human in the middle” job becomes an API call between two models.
6. AI is collapsing the value chain. agencies, recruiters, consultants, and project managers disappear while micro-operators running ten-agent stacks take their place.
7. distribution goes agentic. every AI company will run a thousand influencer agents testing titles, thumbnails, and CTAs nonstop. ad spend becomes a living organism. i hope you like testing.
8. personalization flips commerce. the same product sells for fifty prices through fifty custom funnels, each built by AI for that buyer. price discovery becomes dynamic. this is prob better for business owners and worse for consumers :( .
9. data privacy becomes the new luxury. entire brands form around “human-only,” “no-model,” or “offline verified.” authenticity becomes a trillion-dollar aesthetic.
10. creators will own AI studios instead of channels. one prompt becomes a short, an app, a brand, a product line. the boundary between content and company disappears.
11. the big social platforms fracture into signal markets. people will trade ideas, audience data, and prompt assets the way day-traders swap stocks. virality gets financialized. already happening.
12. energy becomes the next constraint. every AI boom ends in a power bottleneck. whoever solves cheap, local compute with solar or geothermal wins the century.
13. storytelling becomes an economic engine again. the only moats left are narrative, taste, and trust.
14. AI-native insurance becomes a massive opportunity. once agents handle billions of decisions, someone must underwrite the risk.
15. an AI glut means deflation everywhere except in ideas. when intelligence is free, originality becomes priceless.
16. governments create national models to protect sovereignty. data turns into a weapon and compute becomes foreign policy.
17. as agents handle logistics, humans move up the stack into aesthetics. art direction becomes a daily skill. everything becomes branding.
18. the next decade’s wealth comes not just from building AI but from deciding where not to use it. restraint will make fortunes.
19. AI compute arbitrage becomes a trillion-dollar trade. people buy cheap cloud in underdeveloped markets and rent it globally, like Airbnb for GPUs.
20. AI-native brands dominate e-commerce by owning micro-trends. they launch new products daily, test a thousand ad variants, and kill losers overnight.
21. the AI gold rush ends with a massive data rush. whoever owns or licenses niche, verified datasets controls the supply chain of the future.
22. the next $10 billion fund is hybrid: part VC, part compute allocator, part data warehouse. capital moves from money to intelligence.
23. once personal AGIs hit, subscription fatigue dies. consumers will want one AI that handles everything. the first “super-app for life” could be a trillion-dollar company.
24. most billion-dollar outcomes this decade come from repackaging existing industries through AI... the AI accountant, AI real-estate broker, AI logistics coordinator starting as highly vertical versions of familiar services.
25. mobile UI shifts from taps to chat + camera. the screen becomes a lens, the conversation becomes the interface. the app era quietly turns into the agent era. @meetLCA is a design agency i co-founded that is behind the biggest AI apps rn, seeing it play out now.
26. every industry is about to unbundle into interface companies. whoever owns the customer interface, not the backend or the model, controls the value chain. it’s Shopify vs AWS all over again.
27. vertical media merges with vertical SaaS. every niche publication births a product; every software company births a content arm. the media-product line disappears.
28. the internet used to reward consistency. the new internet rewards experimentation. the faster you test, the faster you compound.
29. AI blurs the line between work and art. products start to feel authored, like albums or films. founders become creative directors of automation.
30. AI regulation prob will look like climate policy... too slow, too messy, full of loopholes. innovation moves to places that treat compute like oil.
31. the internet fragments into private ecosystems. niche communities curated by AI become the real web. public feeds feel like Times Square; private groups feel like homes!!
32. the first fully autonomous startup launches within 3 years. no employees, no meetings, no deadlines, just connected agents generating profit. insanity.
33. we are living through the great compression. timelines that used to unfold over decades now happen in months. this is the closest thing to a gold rush most people will ever see.
34. people will look back on 2026–2029 the way we look at the early internet. the difference is you don’t need permission, capital, or credentials. you just need to build something people actually care about.
35. mobile consumer apps feel alive again. they talk back, remember you, and evolve with you. static interfaces begin to feel prehistoric.
36. the next decade of wealth will belong to people who understand three things: distribution is leverage, taste is strategy, and AI is infrastructure.
im tired because i havent slept but wired because...
THIS IS THE BEST TIME IN HISTORY TO BUILD.
our future will look very different than our past/present. life as we know it changing.
i hope you get some sleep.
2
111
Josh Ben-David retweeted
This is the IBIT 2-hr chart going back over the past year. There have been two great buying opportunities in the past year. Guess what? The 3rd is happening as we speak.
This was a massive effort by the team to get Trade launched. There are not many apps out there where you can earn >10% APY on stablecoins and also trade with 40x leverage, in such a simple user experience.
And this is just the beginning. Soon your personal agent will be able to handle all of this for you seamlessly. Stay tuned 👀
🚀 Trading is LIVE on Pass App.
You can now go long or short on top tokens right from your smart wallet.
Trade confidently. Trade smarter. Powered by @HyperliquidX.
👇 Here’s what’s new
3
146
Great advice for builders 💯
This makes no sense. The data here does not tell a clear story.
OP tries to claim that "Big bang TGEs are dead," but half of the "Strong TGEs" are up, half of them are down. And by proportion, most of the successful tokens are "Strong TGEs."
And then the bottom left quadrant of "weak TGEs" only has 8 entries in it, when it should have more like 1000 entries—most TGEs are weak and have negligible hype. That makes it more obvious that Zora and Virtuals are outliers, which is not a strategy.
There is no simple story to tell about how to launch a successful token. Definitely nothing that fits on a Hallmark card. "Launch small and accrue value over time" is another way of saying, "be a sleeper hit," which is not actionable advice.
If people are excited about your token, you cannot launch low. That is a fantasy. If you launch Monad at $500M, it will pop up to multi-billion valuation the next block. You're just giving away money to snipers. It's no different from memecoin launches or NFT mints. If the market believes you're worth X, you're gonna be priced at X, like it or not.
But while I disagree with the analysis of OP, the vibe is clearly dead-on. The rotation of hot money out of the altcoin market and into the equity market over 2025 has challenged a lot of TGEs. But I tend to think these things are all cyclical. People have short memories in crypto, but I remember many points in past cycles where alts were weak, and every time people tell themselves "well that's it, it's finally over, alts will never again catch a bid."
I don't worry too much about that.
But in the meantime, entrepreneurs have to keep building, cycle in, cycle out. My advice to them: focus on product, retain your team, treat your community well, and don't mistake your token price as your report card. Remember what you're building and why you're building it, and trust that if you build something that your users love, the market will ultimately reward you.
Just might take longer than you once thought.
2
110
Josh Ben-David retweeted
I refuse to believe we’re starting a bear market when most altcoin charts look like this.
Bear market start in euphoria
Bull market start in pessimism
They won’t trick me.
Not only a mind bending podcast episode, but he gave notes too 🤩
My pleasure to come on Dwarkesh last week, I thought the questions and conversation were really good.
I re-watched the pod just now too. First of all, yes I know, and I'm sorry that I speak so fast :). It's to my detriment because sometimes my speaking thread out-executes my thinking thread, so I think I botched a few explanations due to that, and sometimes I was also nervous that I'm going too much on a tangent or too deep into something relatively spurious. Anyway, a few notes/pointers:
AGI timelines. My comments on AGI timelines looks to be the most trending part of the early response. This is the "decade of agents" is a reference to this earlier tweet teapawt.org/karpathy/status/188254… Basically my AI timelines are about 5-10X pessimistic w.r.t. what you'll find in your neighborhood SF AI house party or on your twitter timeline, but still quite optimistic w.r.t. a rising tide of AI deniers and skeptics. The apparent conflict is not: imo we simultaneously 1) saw a huge amount of progress in recent years with LLMs while 2) there is still a lot of work remaining (grunt work, integration work, sensors and actuators to the physical world, societal work, safety and security work (jailbreaks, poisoning, etc.)) and also research to get done before we have an entity that you'd prefer to hire over a person for an arbitrary job in the world. I think that overall, 10 years should otherwise be a very bullish timeline for AGI, it's only in contrast to present hype that it doesn't feel that way.
Animals vs Ghosts. My earlier writeup on Sutton's podcast teapawt.org/karpathy/status/197343… . I am suspicious that there is a single simple algorithm you can let loose on the world and it learns everything from scratch. If someone builds such a thing, I will be wrong and it will be the most incredible breakthrough in AI. In my mind, animals are not an example of this at all - they are prepackaged with a ton of intelligence by evolution and the learning they do is quite minimal overall (example: Zebra at birth). Putting our engineering hats on, we're not going to redo evolution. But with LLMs we have stumbled by an alternative approach to "prepackage" a ton of intelligence in a neural network - not by evolution, but by predicting the next token over the internet. This approach leads to a different kind of entity in the intelligence space. Distinct from animals, more like ghosts or spirits. But we can (and should) make them more animal like over time and in some ways that's what a lot of frontier work is about.
On RL. I've critiqued RL a few times already, e.g. teapawt.org/karpathy/status/194443… . First, you're "sucking supervision through a straw", so I think the signal/flop is very bad. RL is also very noisy because a completion might have lots of errors that might get encourages (if you happen to stumble to the right answer), and conversely brilliant insight tokens that might get discouraged (if you happen to screw up later). Process supervision and LLM judges have issues too. I think we'll see alternative learning paradigms. I am long "agentic interaction" but short "reinforcement learning" teapawt.org/karpathy/status/196080…. I've seen a number of papers pop up recently that are imo barking up the right tree along the lines of what I called "system prompt learning" teapawt.org/karpathy/status/192136… , but I think there is also a gap between ideas on arxiv and actual, at scale implementation at an LLM frontier lab that works in a general way. I am overall quite optimistic that we'll see good progress on this dimension of remaining work quite soon, and e.g. I'd even say ChatGPT memory and so on are primordial deployed examples of new learning paradigms.
Cognitive core. My earlier post on "cognitive core": teapawt.org/karpathy/status/193862… , the idea of stripping down LLMs, of making it harder for them to memorize, or actively stripping away their memory, to make them better at generalization. Otherwise they lean too hard on what they've memorized. Humans can't memorize so easily, which now looks more like a feature than a bug by contrast. Maybe the inability to memorize is a kind of regularization. Also my post from a while back on how the trend in model size is "backwards" and why "the models have to first get larger before they can get smaller" teapawt.org/karpathy/status/181403…
Time travel to Yann LeCun 1989. This is the post that I did a very hasty/bad job of describing on the pod: teapawt.org/karpathy/status/150339… . Basically - how much could you improve Yann LeCun's results with the knowledge of 33 years of algorithmic progress? How constrained were the results by each of algorithms, data, and compute? Case study there of.
nanochat. My end-to-end implementation of the ChatGPT training/inference pipeline (the bare essentials) teapawt.org/karpathy/status/197775…
On LLM agents. My critique of the industry is more in overshooting the tooling w.r.t. present capability. I live in what I view as an intermediate world where I want to collaborate with LLMs and where our pros/cons are matched up. The industry lives in a future where fully autonomous entities collaborate in parallel to write all the code and humans are useless. For example, I don't want an Agent that goes off for 20 minutes and comes back with 1,000 lines of code. I certainly don't feel ready to supervise a team of 10 of them. I'd like to go in chunks that I can keep in my head, where an LLM explains the code that it is writing. I'd like it to prove to me that what it did is correct, I want it to pull the API docs and show me that it used things correctly. I want it to make fewer assumptions and ask/collaborate with me when not sure about something. I want to learn along the way and become better as a programmer, not just get served mountains of code that I'm told works. I just think the tools should be more realistic w.r.t. their capability and how they fit into the industry today, and I fear that if this isn't done well we might end up with mountains of slop accumulating across software, and an increase in vulnerabilities, security breaches and etc. teapawt.org/karpathy/status/191558…
Job automation. How the radiologists are doing great teapawt.org/karpathy/status/197122… and what jobs are more susceptible to automation and why.
Physics. Children should learn physics in early education not because they go on to do physics, but because it is the subject that best boots up a brain. Physicists are the intellectual embryonic stem cell teapawt.org/karpathy/status/192969… I have a longer post that has been half-written in my drafts for ~year, which I hope to finish soon.
Thanks again Dwarkesh for having me over!
185
Ever thought you're earning “passive yield” only to realize impermanent loss ate half your profit?
@yieldbasis finally solves this. It turns LPing into an automated, adaptive strategy.
It watches prices, adjusts positions, and converts volatility into income in real time.
Now you no longer have to hope yield beats impermanent loss.
Huge shift in the LPing model. Hope to see more assets supported over time.
155

























