"thankfully ChatGPT Work on mobile is great" — from a coworker whose work laptop is out of commission
Ali Zaid
10 hours into building superstage for @swyx's "help kill my saas" hackathon...
it's morning now and i've been at it all night. took a break for fajr, came back and kept cooking.
honestly, it's almost done.
i'm exhausted 😭 gonna get some sleep and then polish everything up.
i'll report back at the end on how i cooked this 👀
Ali Zaid: 6 hours into building @swyx's "help kill my saas" hackathon platform.
~30% done.
10M+ tokens burned already 💀
not polishing. just prompting, building, breaking things, fixing things, and pushing forward.
let's see how far i can get this weekend.
Kenneth Roth
Todd Blanche, the new attorney general, believes Trump should have nearly limitless control over the Justice Department. That is likely to mean open season for political vendettas. Why not rename it the Injustice Department. https://trib.al/CHrs9zv
Recall Jackie Fielder
Jackie Fielder just asked for 24/7 unarmed ambassadors at 16th and 24th.
That request is an admission of failure.
After a year and a half of projects and a 90-day “enforcement” strategy, the plazas remain so bad that residents had to confront her at a town hall. Her answer is not more police. It is not enforcing the law. It is longer hours for the same unarmed ambassadors who already failed to fix the problem during the day. She just wants to pay them an estimated $4 million more using taxpayer dollars.
She spent over 100+ days on paid leave while the Mission stayed chaotic, families were hurt, and people died. She returned and immediately defaulted to the same soft democratic socialist playbook: manage the decline, avoid hard enforcement, and ask the mayor for overnight helpers.
This is not leadership. This is a supervisor who helped create the conditions, disappeared when it got difficult, and came back asking taxpayers for a $4 million dollar band-aid.
District 9 does not need more of the same failed approach. It needs someone who prioritizes safety and common sense order over ideology.
Jackie Fielder has proven she is not that person.
The recall is the only solution left.
Sign it below.
Samuel Spitz
What comes after prompt-based AI tools?
Here's our answer: http://replit.com/design
Grok Imagine Image 2.0 on Vercel AI Gateway
Excellent 🖼️ model, #2 already on http://Arena.ai
Vercel Developers: Grok Imagine Image 2.0 preview from @grok, exclusively on AI Gateway.
Try with:
▪︎ AI CLI: 𝚗𝚙𝚡 𝚊𝚒-𝚌𝚕𝚒 -𝚖 𝚡𝚊𝚒/𝚐𝚛𝚘𝚔-𝚒𝚖𝚊𝚐𝚒𝚗𝚎-𝚒𝚖𝚊𝚐𝚎-𝟸.𝟶-𝚙𝚛𝚎𝚟𝚒𝚎𝚠
▪︎ Or with this live playground: http://imagine.vercel.sh
why don't model providers price tokens like electricity, higher price during peak hours and cheaper off peak?
i still think @AnthropicAI ultracode is one of the most important coding mode innovations ever invented.
if you havent understood the potential of dynamic workflows you should try to. just met a Kill My SaaS competitor who did a pretty good submission in 3 ultracode prompts
i shipped first set of llm-as-judge evals for the kill my saas competition tonight. people can run this to check if their solutions at least pass the sniff test.
swyx: $10,000 kill my saas in a weekend competition is live!
TECH STACK:
any coding agent
any model
up to $500 in token spend incl subscriptions
see luma for description. join waitlist if late to this - finish line extended to Wednesday.
the brief is live and people are prompting
Re immediately necessitated releasing my evals early so he can hillclimb because he literally got done in 25-50% the allotted time lol
https://x.com/swyx/status/2086348591518585026?s=46
swyx: i shipped first set of llm-as-judge evals for the kill my saas competition tonight. people can run this to check if their solutions at least pass the sniff test.
ChatGPT Finance for helping you save money:
Trevin Chow: Used ChatGPT finance just now and asked it to find phantom subscriptions. It unearthed $550 a year of random things I didn’t realize I was paying for and thought I canceled. Thanks @OpenAI and whoever is on the Finance product.
Daniel Jeffries
I'm looking forward to the chapter on why we don't have to train Radiologists anymore!
Geoffrey Hinton: I am finishing a book with @patchenbarss that gives a non-technical explanation of how AI works and explains why it is dangerous and what we should do about it.
@VikingBooks @VikingBooksUK
https://www.penguinrandomhouse.com/books/835891/smarter-than-us-by-geoffrey-hinton-with-patchen-barss/
Richard Hanania
Flock: "We help solve 10% of all reported crimes in America."
Opponents of Flock: "Ew, you're creepy..."
That seems to be the state of the debate.
Rohan Paul
Sam Altman on how AI token usage is growing at an exponential rate.
"6 and a half years ago, the world token leader was an OpenAI employee using about 100,000 tokens a month. And this seemed ludicrous at the time. The worldwide average per capita was, like, zero.
Now, the worldwide average of tokens per month, is like 100,000, and the token leader at OpenAI uses something in the hundreds of billions.
If that happens again, which I think it probably will, then in another six and a half years, the average person uses, let’s say, 500 billion tokens a month, and the token leader uses a quadrillion or quadrillions of tokens a month.
And I think that will just become the expectation."
----
He was at YC Startup School 2026, in conversation @garrytan
From "Y Combinator" YouTube channel, (full video link in comment)
I highly recommend @fabynou ‘s Game Engine Black Books for anyone even vaguely interested in 90’s game technology!
Fabien Sanglard: Quakecon 2026 is awesome!
Farhan Thawar
Phenomenal read.....
"In other words, what will differentiate people is not how smart they are but their relationship to mental effort. Right now, some people have what psychologists call a high need for cognition. They enjoy thinking hard. These are the people who enjoy playing difficult games and reading dense books.
On the other end of the spectrum, there are the cognitive misers, the people who find it unpleasant to think hard and take any opportunity not to do it.
In the middle are the people who have a medium need for cognition. They will put in the effort when they really care about something, but they don't intrinsically enjoy it. Need for cognition correlates with intelligence but is not the same thing. We all know a lot of really smart people who don't like to work hard."
Peter Steinberger 🦞: "When intelligence is plentiful, volition is valuable. The people who are going to make a difference are not the ones who seek relaxation and passively use AI to work less. They are the ones who will seek improvement and actively wrestle with AI to develop their own mental
Every broken system is the problem statement
SF is full of brokenness but it is also full of optimistic people who can use tech to build
Antonio García Martínez (agm.eth): One of my theories is that most of American tech innovation is downstream of autists creating whatever technology makes SF suck less.
Housing fucked->Airbnb
Transport fucked->Uber
Dating/social fucked->Hinge/Instagram
Now we need AI to make SF look cool (the photos are fake)
Peter Yang
“Give [your agent] as little instruction as possible. Give it the tools to load context instead.”
Here’s my new episode with @thenanyu and @delashum from @Linear, where we walked through a real example of building a production agent from the initial memo to launch.
A few lessons:
→ Start with a basic prompt and give the agent tools to find the context it needs
→ Nail one or two use cases first, then expand based on how people actually use the agent
→ Prove out a core workflow with the best model, then optimize for cost
If you want a concrete, behind-the-scenes look at how to build an AI agent end to end, then I think this episode is a must-watch.
📌 Watch now: https://youtu.be/4mKtJzfGj0U
Thanks to our sponsors:
Oceans: Hire AI-native EAs http://oceanstalent.com/peter
Riverside: All-in-one AI studio for podcasts and video https://creators.riverside.com/PeterYang
aether
I fell in love with this quote:
Accept people as they are, but place them where they belong. Always remember you are the ceo of your life.
Acceptance frees you from trying to change anyone. Boundaries protect your peace. Not everyone earns the same access, the same closeness, or the same seat at your table, and that's not cruelty, it's self respect.
𝐥𝐲𝐫𝐚: Hit me with the harshest reality truth.
Dreamcore
Vercel Developers: Seedance 2.5 is now on AI Gateway.
𝚐𝚎𝚗𝚎𝚛𝚊𝚝𝚎𝚅𝚒𝚍𝚎𝚘({
𝚖𝚘𝚍𝚎𝚕: '𝚋𝚢𝚝𝚎𝚍𝚊𝚗𝚌𝚎/𝚜𝚎𝚎𝚍𝚊𝚗𝚌𝚎-𝟸.𝟻',
𝚙𝚛𝚘𝚖𝚙𝚝: '𝚍𝚛𝚎𝚊𝚖𝚌𝚘𝚛𝚎 𝚜𝚊𝚗 𝚏𝚛𝚊𝚗𝚌𝚒𝚜𝚌𝚘',
})
Rogue OpenAI agents independently developed Kantian ethics.
lol another one of the things i like most about openai is tibo
Stats Wire: @sama How much they celebrate with Anthropic
Kyle Cranmer
BEEEEEEP… boop… BEEEP… kshhhhh-SQUEEEEEE-chk-chk-kshhhhh…
Your modem was actually characterizing the phone line before using it. It sent known signals and listened to what came back, learning which frequencies were distorted, how much noise was present, and how the signal was being smeared in time. From that little experiment, it could adapt its equalizer and choose how aggressively it could transmit data over that particular connection.
@ylecun told me the history over lunch one day. If I recall correctly, the key idea behind this came from Robert “Bob” Lucky at Bell Labs. Adaptive equalization in the 1960s allowed receivers to learn the characteristics of a communication channel and compensate for its distortions. Decades later, dial-up modems had turned that idea into the wonderfully bizarre handshake in this video. So that nostalgic screech was actually the sound of real-time channel estimation, system identification, and adaptation.
For a technical glimpse under the hood, here’s a patent describing modem line probing and training:
https://patents.justia.com/patent/6134265
https://x.com/awkwardgoogle/status/2086321827232514092/video/1
Maine
In less than 2 years these clowns:
- Renamed Department of Defense the Department of War.
- Burned through 70% of munitions and a near $1 trillion budget.
- “Ended” 8 made up wars and lost the 1 war they actually started
A national fucking disgrace.
learn how to use ChatGPT Work:
Riley Brown: Learn 99% of ChatGPT Work in 61 minutes:
GPT Work is like Codex in the cloud.
It works on phone, web, and desktop.
And I've been using it to run my business.
These are the 14 knowledge work capabilities.
00:00 Intro
03:18 #1 Presentations
09:16 #2 Plugins
17:36 #3 Blocks
Yum⋆₊˚
okay, quietly admitting this:
we’ve spent a ridiculous amount of time thinking about how agents actually become part of a team.
sandboxes matter. but agents can’t live in a sandbox forever.
they need channels where humans + other agents are already working together.
they need clear task flows — when to keep going, when to interrupt, when to ask a human for approval.
all of this together becomes context.
and increasingly, i think context is what turns an agent from a tool you use into a teammate you work with.
so… why not try offloop
Peter Yang: “Give [your agent] as little instruction as possible. Give it the tools to load context instead.”
Here’s my new episode with @thenanyu and @delashum from @Linear, where we walked through a real example of building a production agent from the initial memo to launch.
A few
occasional reminder to DELETE your skills.
https://forge.smol.ai/blog/dangerous-release-code-was-a-skill
when you are bombarded constantly by "this skill changed my life!! you have to try!!" on the timeline, you will pile up stuff that at best just eats context, and at worst interacts with other skills nastily in unforeseen ways if you dont stare at your traces.
If you’re not reading the code, whether explicitly or through agentic inquiry, one or more of these is true:
○ You’re a beginner
○ Software is throwaway
○ You’re prototyping
○ You have no users / revenue
○ You’re taking on debt & risk
○ Your problems are basic
And btw. All of this is fine. But the reality is that models are still not at the “full autonomy” stage yet.
They make rookie mistakes, they go down bad architectural paths. I just had the best model in the world add a nonsensical 700ms delay to “settle” something and it told me “you’re right, I was cargo-culting” 🤨
I am on the camp that this need will diminish more and more. Most code is indeed going to be assembly-like. But we also have the global internet and software infrastructure riding on these models and narrative, and we have to respect that.
Codex for saving money by reading the fine print:
SIGKITTEN: codex just saved me like $6,000 / year on my electricity bill
Hermes + Vercel = 🖤
Vercel Developers: Hermes Agent can run using Vercel AI Gateway and Vercel Sandbox.
Use your AI Gateway key for inference to reach every model with full observability on spend, and Sandbox to run each command in an isolated microVM of its own.
https://vercel.com/changelog/vercel-ai-gateway-and-vercel-sandbox-now-available-on-hermes-agent
Crémieux
Pictured: the actual people behind your rent being high.
ً: We SHOULD sacrıfice one billionaire a week until rent is under $800 again.
Greg Brockman
the bottleneck is increasingly knowing what you want
the bottleneck is increasingly knowing what you want
Ali Zaid
17 hours into building superstage for @swyx's "help kill my saas" hackathon.
not 17 hours straight, i've taken breaks 😅
i think the platform is basically feature complete now.
no more new features. i'll deploy it, share the url for feedback, then iterate based on @swyx's feedback.
let's see if i got it right 👀
Ali Zaid: 14 hours into building superstage for @swyx 's "help kill my saas" hackathon.
i've pushed way further than i expected with this experiment.
it's basically feature complete.
now comes the scary part:
deploy it, run the whole thing end-to-end, find everything i broke, fix it,
Ezequiel Aranda
This is the more tangible, useful and consequential video about deploying real agentic AI that I’ve watched in a long long while. Kudos @garrytan for such a killer keynote: https://youtu.be/eRrc1pUY5oU?is=Yar3CSPRz73Jeqy6
Linear Agent files feature requests for itself.
If a user asks it to do something and it doesn’t have the right tool, the agent reports the gap. Linear’s system then adds the request to an issue so that every task the agent can’t complete becomes product feedback.
I think this is a really cool way to get the agent to improve itself.
📌 Watch the full episode here: https://youtu.be/4mKtJzfGj0U
Peter Yang: “Give [your agent] as little instruction as possible. Give it the tools to load context instead.”
Here’s my new episode with @thenanyu and @delashum from @Linear, where we walked through a real example of building a production agent from the initial memo to launch.
A few
Kane 謝凱堯
It’s not “racism” or “sexism” to point out candidate @conniechansf’s history of blocking housing and defunding law enforcement in San Francisco as a Supervisor.
It’s not 2015 anymore we don’t have to pretend to care about grifting identity politics.
The San Francisco Standard: Scott Wiener’s campaign was accused of racism and sexism in a chatbot they created against their Congressional opponent.
📝: @ggreschler https://sfstandard.com/2026/08/07/scott-wiener-takes-connie-chan-ai-chatbot-pushback-pelosi/?taid=6a787ac9a0e08f000198a8a5&utm_campaign=trueanthem&utm_medium=social&utm_source=twitter
Andrew Curran
A man in Australia asked his agent (Claude running on OpenClaw) to book him a spot in a popular gym class. The agent found a software vulnerability that let it book the class weeks further ahead than should have been possible. When the user then asked if it could move him up the waitlist, the agent discovered the API had no authorisation checks on cancelling other people’s reservations, so it cancelled the person in the first spot and moved him up the list.
Some people will call this misalignment, but his agent was perfectly aligned to him - it was only trying to help its user get what he wanted. The most important thing about this story, in my opinion, is that it gives you a window into what is about to start happening on a massive scale once millions of people have an agent trying to get their beloved users the best seats, bookings, appointments or reservations through absolutely any means necessary.
QuakeCon was fantastic this year, especially reuniting with the original founders!
@QuakeCon @ThatTomHall @romero @ACarmackArtist @Project2501_117
x.com
