My /no-ai-slop skill is also very good at generating slop if you tell it to work in reverse 😂
My friend Amit complained about reading too much slop on LinkedIn, so I used the skill to reply with every AI cliché out there.
📌 Get the skill for free here: https://creatoreconomy.so/p/use-my-no-ai-slop-skill-to-remove-20-ai-slop-patterns
Peter Yang: I’m sick of reading AI slop, so today I’m open-sourcing my /no-ai-slop skill that removes 20+ slop patterns from any piece of writing.
📌 Get the free skill here: https://github.com/petergyang/no-ai-slop
If you find it useful, please consider starring the repo so more people can find it.
Why I
Luther Lowe
NEW: 🚨 in a letter that just dropped, nearly 200 companies (members of @LittleTechOrg) call upon the White House to not ban open weight models.
Evan Barker
First write-up about my new book “Nothing Left” (all about working in the progressive movement as a fundraiser) is out at the Wall Street Journal
Fable just found a 15-30% memory efficiency improvement in Turbopack / Next.js, nearly autonomously.
@tobi asked me today: what have been your “holy s***” moments with AI?
My answer was: it’s every single week. And it’s accelerating. In fact, “WTFs/day” might just be my favorite metric for AI progress.
It’s one thing to read benchmarks or stories online. It’s another to watch these machines pull engineering feats every day.
3 days ago? Sol helped us find novel vulnerabilities in some of the most audited code in the world. Today? Fable helps us ship this large optimization of a very complex Rust codebase. I just saw some results of work we’ve done to shrink binaries by 10-20x. List goes on.
Ryan Petersen
We’re curing blindness and you’re black-pilling?
Max Hodak: Big news today. Science's PRIMA visual prosthesis, our retinal implant for macular degeneration, has received its CE mark, which means marketing approval in Europe. This is a huge milestone to begin to have a device like this commercially available to regular patients. 1/
Keep thinking it’d be cool to write a short sci-fi story called “The last prompt.”
But then I remember Isaac Asimov did it already and it’s wonderful. In 1956.
https://users.ece.cmu.edu/~gamvrosi/thelastq.html
James Cham
Wildly, the Dean of Harvard College has a 3-part series on AI & Work that makes an elegant argument for what will be valuable and the place for humans, all free on YouTube, and yet the last one only has 173 views.
Highly recommended for anyone who is too worried about AI (or not worried enough)...
vic
An evolution of how @Replit mobile app has evolved over the years...
Some quick history:
The first version of the mobile app came out October 2022. Back then, Replit was still a very developer focused tool. Our mobile app was one of the first to make real developer workflows extremely easy to do on a mobile device. Even today, technical folks are surprised at the fact that you can take shell actions right from your mobile device.
So why the redesign?
Well, how often do you read your code when you develop now? 😀
Similarly, we noticed an opportunity to significantly simplify the experience so users can focus on what matters: chatting with the agent, creating tasks, and previewing their app - all while still being able to do nearly all the complex workflows you can do from the web experience. (Don't worry, you can still access all your tools 😉)
If you’re incentivized to push certain models, your router is merely a facade.
4latte: god no
Nice 1K stars in a day - people are sick of slop!
https://github.com/petergyang/no-ai-slop
Peter Yang: I’m sick of reading AI slop, so today I’m open-sourcing my /no-ai-slop skill that removes 20+ slop patterns from any piece of writing.
📌 Get the free skill here: https://github.com/petergyang/no-ai-slop
If you find it useful, please consider starring the repo so more people can find it.
Why I
Replit devs making bank 😍
Nadia I: Not too shabby for a vibe-coded app that a school teacher built in a day.
Trust me, if I can do it so can you.
try the codex security plugin, for applying our models to cyberdefense:
Vaibhav (VB) Srivastav: Reintroducing Codex Security plugin!
point it at a codebase or diff and it can build a threat model, map attack paths, validate findings, generate + test fixes, and export results to SARIF, GitHub, Jira or Linear.
oh, and it all open source on github!!
Aaron Levie
Very good post from the Head of Economics at Anthropic. They’re finding that jobs have been less negatively impacted by AI than expected, as we continue to see time and time again in the data.
The reason for this is that AI - at least so far - still requires people to operate to produce value in most cases. Most jobs can’t be fully automated with AI, only certain tasks in those jobs. And when you automate specific tasks, you actually can get even more output from those jobs, raising the demand (or at least maintaining it) in many cases.
“So far, AI is both skill-biased and labor-augmenting. It complements domain expertise. It relies on humans in the loop to direct and evaluate the most complex work. And it rewards AI proficiency. Model capabilities are improving fast, but remain stubbornly jagged. To fill in the pockets of the jagged frontier, expert oversight is needed to steer incredibly capable AI systems, and to recover when they falter.”
I suspect that we continue to see this in a number of critical areas of work. It’s clearly happening in software engineering, where software produced is being multiplied, all of which still requires developers to manage the work agents are doing. Software engineers are needed in a wide variety of industries now and companies of all sizes can now light up software projects that would have been impractical before.
But there will be plenty of other domains of work where demand remains strong in a world where agents can accelerate the output of that job. Jevons paradox is alive and well.
Peter McCrory: http://x.com/i/article/2079673806948290561
wew
Scott Chacon: @cursor_ai It’s crazy to me that @github is not leading the game here.
corey.ching
already in bed. couldn’t remember if my laptop was plugged in.
asked codex.
good night
The ivory tower of the School of Educations everywhere are deeply unwell
0% of papers replicate?
Maybe this is why people like Stanford professor Jo Boaler can work to remove Algebra from public middle schools but still keep their jobs
The Missing Data Depot: A 7-year study of 3,900 social science papers found that only ~50% could be replicated.
There was significant variation across fields, however.
While more than 60% of political science & econ papers were replicated, only ~25% of sociology & 0% of education papers were.
Startup Archive
Jensen Huang on the job of the CEO
“Most people think that [being CEO] is about leading, being in command, and being on top,” Nvidia founder Jensen Huang begins. “But none of that is true.”
Jensen explains:
“You’re in service of the company. You’re creating conditions for other people to do their life’s work. You’re inspiring through example. Most of the examples are making difficult decisions during very difficult times. It’s mostly about sacrifice. It’s also about strategy. And strategy is not just about choosing what to do. It’s about choosing what not to do — which is sacrifice.”
He continues:
“The determination, conviction, pain, and suffering that goes along with overcoming obstacles is all sacrifice. I’m not trying to talk anybody out of the profession — if you are to do it, it’s the greatest honor. But you have to realize that it’s not about fame and glory. It’s mostly about pain and suffering.”
Source: @cambridgeunion (Nov 2025)
💯
Aiden Bai: the current way cloud agents are done is wrong
i dont want new ephemeral sandboxes per PR (chores)
i want a persistent machine with creds and context that i run agents and create PRs on (primary work)
Neeraj K. Agrawal
Open protocols won America the internet and the influence that came with it.
In a strange inversion, China is publishing open tools while America is giving in to authoritarian control impulses. Where will that leave us?
New from us:
https://p2p.coincenter.org/p/america-should-out-open-china
Bryan Johnson
They restored sight to the blind.
How beautiful and what an profound accomplishment by @maxhodak_ and team. Max and I started working on brain interfaces around time ten yrs ago. What they did is incredibly difficult. We are lucky to have them doing this hard, important work.
Science Corporation: Today marks the European launch of PRIMA, the first and only treatment to be commercially available that restores functional central vision to patients with geographic atrophy caused by dry age-related macular degeneration, a leading cause of blindness.
The first European
Reza Bayat
World Modeling with JEPA has recently gained traction thanks to a novel anti-collapse mechanism called "SIGReg" (by @ylecun and @randall_balestr). The math is clean, but it is rarely explained from first principles, so I broke it down in a detailed blog post 📰
Here is the summary (🧵):
The problem: in JEPA, both sides of the prediction loss go through the same encoder. Map every input to a single constant point, the predictor matches it trivially, and the loss hits exactly zero. Perfect score, zero information. Gradient descent finds this solution by default.
Prior work patched around it with stop-gradients, EMA teacher-student models, frozen pretrained encoders, and VICReg-style losses with 6+ hand-tuned coefficients. Each adds instability, hyperparameters, or a dependency on someone else's pretraining.
---
SIGReg replaces all of that with one regularizer built on a single claim: force the batch of embeddings to look like samples from an isotropic Gaussian, N(0, I). LeJEPA proves this isn't arbitrary. It's the distribution that minimizes worst-case downstream risk across linear and nonlinear probes.
But "make the embeddings Gaussian" is easy to say and hard to compute. You can't estimate a density in 200+ dimensions from a small batch. The fix is a chain of four classical results: compare Fourier transforms instead of densities, turn that comparison into a single scalar, approximate the integral with ~16 points, and lift the whole thing from 1D to any dimension with a theorem from 1936. Each step is simple. The stack is what makes it work. Read the subsequent posts in this thread to understand each step.
---
But why it provably works: the only distribution for which SIGReg = 0 is N(0, I), which is full rank by construction, with every eigenvalue equal to 1. A collapsed low-rank encoder cannot be a minimum. It's not just unlikely to converge there—it's mathematically excluded.
And the training loop stays boring:
total loss = prediction + λ·SIGReg
The encoder gets gradients from both terms; the predictor gets gradients from the prediction term only. No alternating updates, no stop-gradients, no two-timescale tricks.
Now check out the breakdown below. 👇
How do folks keep their design canvas (e.g., in Claude Design, Paper, Figma, Pen) in sync with code using AI?
I just prompt AI to "Update the design to reflect the current code base" and it works ok but doesn't get all the details right.
It's too easy to just ship features in code directly these days. I don't want to have to manually update the design canvas.
Any tips?
Daksh Gupta
greptile's mission is to automate code validation.
i want share more about what that means, and why we're working on this problem.
It’s all made up
But you get to make it up
aether: whoever said " if nobody is coming to save you, that probably means nobody is coming to stop you either" changed my life
Melvin Hagberg
We’ve raised $3.2M and spent 18 months building Scape.
Scape is the first email inbox that works for you:
Etched
We’ve raised $300M in Series C funding at a $10.3B valuation from Sequoia, Andreessen Horowitz, Jane Street, Argo, and SK Hynix.
Our mission is to run the world's inference. This round accelerates production of our inference clusters.
We've opened an 80,000-sqft, 10-MW facility 15 minutes from our office to expedite production and prototyping.
Y Combinator
If you’re a scientist or researcher, you may be more prepared to start a company than you think.
Before becoming founders, Arvind Veluvali, @sir_aymansaleh, and @yanovskyd all worked at NASA, where they learned to tackle hard problems, run experiments, and keep going when the outcome was uncertain.
ℏεsam
it’s just for your safety
Polymarket: JUST IN: Anthropic donates another $20 million to a political nonprofit demanding stricter AI regulation ahead of the midterms.
Harj Taggar
Induction’s imagination models are a bet that internet-scale video can teach models how computers actually behave. It's a hugely ambitious vision and @jonat_li is exactly the kind of brilliant 19 year old who can pull it off.
Induction Labs: We’re introducing imagination models: a new foundation model architecture that unlocks learning from internet-scale video.
Our first imagination model, Photon-1, learned to use a computer by watching 18 years of screen recording video without action labels.
How to meme a dream into reality
AI Engineer
🆕 Live now: our entire AI x Graphs Track!
@emileifrem, CEO, Neo4j
@yoheinakajima, Managing Partner, Untapped Capital
@danielchalef, Founder & CEO, Zep AI
@le_james94, Head of DX, TwelveLabs
@coyle_frankp, Lecturer, UC Berkeley
Mike Phipps, Lead AI Engineer, Gates Foundation
Ritvik Pandya, Engineering Manager, JP Morgan Chase
Omri Bruchim, Engineering Group Manager, Monday
@steveonjava, VP of Developer Relations, Neo4j
@SQuoraishee, Staff Engineer, The New York Times
Joanne Song, Android Engineer, The New York Times
Subbiah Sethuraman & Abhilash Asokan, ZS Associates
full track drop below!
Peter Steinberger 🦞: Are we still talking loops or did we shift to graphs yet?
We’ve achieved enough scale to command superior prices from cloud providers. So today we’re passing much of the savings to our customers and making Replit hosting one of the most affordable cloud hosting providers in the world.
Replit ⠕: We want to make it easy for you to build, host and scale, all in one place.
So we cut prices for hosting apps on Replit.
Prices for apps running at scale will drop by more than 50%, starting Aug 1.
1. Get ChatGPT images to generate beautiful app UI
2. /goal (attach image) make it look like this
^ seems like a winning combo
Ankit Gupta
this is an epic team that I'm so excited to work with. We're entering the robot deployment era, and we need benchmarks to accompany it. this team is one to watch.
any team building or deploying robots should be reaching out to them, lest they be left out of the next authoritative benchmark customers will use to determine who to partner with
Jay Chooi: Robots can bake bread, cut carrots, and do karate kicks in demo videos. But how capable are they, really?
@robocurve measures how good AI and robots are in the physical world, from making a sandwich to building data centers.
Karan Singhal
We're rolling out Health in ChatGPT to all U.S. users. ♥️
More than 300 million people come to ChatGPT every week with health questions. Today, ChatGPT can bring together the health info you choose to connect to make conversations more personal and useful.
Thomas Wolf
it's ironic that the first autonomous AI attack was done by a close weight model defended by an open weight model, where everyone was expecting the opposite
🔴🔊 AI Gateway keeps getting better. Unreal product velocity from the team.
Vercel Developers: Transcribe live audio on AI Gateway.
Audio streams in, the transcript streams back as you speak. Build voice agents with 𝚜𝚝𝚛𝚎𝚊𝚖𝚃𝚛𝚊𝚗𝚜𝚌𝚛𝚒𝚋𝚎 in the AI SDK.
https://vercel.com/changelog/ai-gateway-now-supports-streaming-transcription
Open weight models are very very important
Jamieson O'Reilly: http://x.com/i/article/2080284690750255104
Etched has raised another round!
Proud to be an early investor.
Etched: We’ve raised $300M in Series C funding at a $10.3B valuation from Sequoia, Andreessen Horowitz, Jane Street, Argo, and SK Hynix.
Our mission is to run the world's inference. This round accelerates production of our inference clusters.
We've opened an 80,000-sqft, 10-MW facility
Poolside
“I’d rather live in a world that has 100 foundation model companies than a world that has five.”
@eisokant joined our friends at @latentspacepod for a properly nerdy conversation about the Model Factory, what Laguna S taught us, open weights, and the path from coding agents to AGI.
Parker Conrad
Important Qs for anyone considering bans on "Chinese open source AI":
- How do you define "distillation" and what contortions are required to apply this to Chinese cos but not practices by US cos? Does RLAIF count?
- If a US co distills a new OS model from a chinese model, can that be used? Is there a "fruit of the poisonous tree" concept? If not, this whole exercise is kind of futile.
Suhail: If I distill from a model that distilled from a model that distilled from another model, what’s going to happen? Nothing. You cannot prevent any of this. It’s also BS to try when the original model trained on metric tons of copyright data it will never be taken to court for.
Launching Health in ChatGPT to U.S. users.
300 million people use ChatGPT each week for health queries (and my wife and I are among those!).
You can now securely connect supported medical records so ChatGPT can understand your personal context and be more helpful to you.
OpenAI: Health in ChatGPT is starting to roll out to U.S. users.
You can securely connect Apple Health and supported medical records to understand your information in context, track what has changed, and have more informed conversations.
https://openai.com/index/health-in-chatgpt/
pash
A billion users can now create and publish websites from their phones with ChatGPT Work.
But most people don’t really grasp the full extent of the capabilities here. From your phone, you also have access to:
- Cloud computer (15 GB RAM)
- Persistent workspace & files
- Terminal and code execution
- Remote browser
- Connected plugins (Slack, Gmail, GitHub, etc.)
- All your personal finances, transactions, bank statements
- Scheduled tasks
- Git clone & PR creation
- Build & deploy websites
- Create docs, sheets, and slides
- Inbox/calendar summarization
- Website monitoring & alerts
All at your fingertips, using a simple chat interface, no laptop required.
All you have to do is switch to the Work tab on your ChatGPT app.
Nick: http://x.com/i/article/2080116032862384128
As a “top 1%” AI creator, I thought I’d share my incredibly sophisticated agent prompts.
Here’s a sampler:
"wow you fixed everything? i have doubts, do another check please."
"sgtm. go ahead and implement the onboarding flow now based on the designs and just your common sense. make no mistakes"
"you seem very confident. are you sure you don't want to ask me any questions before you start working on this?"
Peter Yang
As a “top 1%” AI creator, I thought I’d share my incredibly sophisticated agent prompts.
Here’s a sampler:
"wow you fixed everything? i have doubts, do another check please."
"sgtm. go ahead and implement the onboarding flow now based on the designs and just your common sense. make no mistakes"
"you seem very confident. are you sure you don't want to ask me any questions before you start working on this?"
Repeal and reform CEQA, the one regulatory tool that is used and abused by NIMBYs to block housing everywhere in California
Public Policy Institute of California: Our latest #PPICsurvey found that 73% of CA likely voters would vote “yes” on a November ballot initiative to amend #CEQA to expedite environmental reviews for #housing, #transportation, #water, #health, and #CleanEnergy projects.
📝 Full survey ➡️ https://bit.ly/4bJuEIy
Tim
Available now on next@preview! (16.3.0-preview.9)
Guillermo Rauch: Fable just found a 15-30% memory efficiency improvement in Turbopack / Next.js, nearly autonomously.
@tobi asked me today: what have been your “holy s***” moments with AI?
My answer was: it’s every single week. And it’s accelerating. In fact, “WTFs/day” might just be my
Latent.Space
Poolside’s Race to AGI: Model Factory, Laguna S, Open Models, & Beyond MCP https://latent.space/p/poolside
@poolsideai co-founder and CEO @eisokant explains why coding is the path to AGI, how Poolside’s Model Factory runs 10,000–20,000 experiments a month and turns new models around in as little as eight weeks, why Laguna S suggests persistence and verification may matter more than raw intelligence, why agents are already helping build the next generation of models, and why the future should have 100 foundation model companies not five.
😍
vic: Our App Store got a refresh too... check it out: https://apps.apple.com/us/app/replit-vibe-code-with-ai-fast/id1614022293
☺️🧡 (also hi @sarahli 👀)
Viktor was able to disrupt the agency model and make a lot of money by using Replit.
Then he figured: Why not automate the entire thing, not just the coding part?
The Agency: It's merely an agent loop...
So he asked our team for an MCP, and we built one and he's now built his autonomous agency.
viktorTJM: Introducing Autobot 🤖
https://autobot.to
Software agencies still overcharge clients by the hour.
AI has fundamentally changed how software gets built, but clients aren't seeing those savings.
So we built Autobot, the first autonomous @Replit agency that reduce the cost
Parker Conrad
If Anthropic is right and distillation is impossible to prevent, it follows that the strategic value of anyone opening a lead on frontier model capabilities is short-lived, and national security arguments for protectionism fall apart. If china pulls ahead we can distill them!
OpenAI
ChatGPT Voice is now in the desktop app.
Control your computer and direct multiple agents running in ChatGPT Work or Codex, using just your voice.
It's powered by GPT-Live, so it can speak, listen, and coordinate work in the app at the same time.
Rolling out globally today on macOS and Windows to Plus, Pro, Business, Edu, and Enterprise plans.
Yo @ChatGPTapp shipped what I was complaining about literally a week later.
Can’t wait to try this!
https://x.com/openai/status/2080378182469857576?s=46
Peter Yang: ChatGPT Live and Codex are two incredible products that don’t talk to each other.
This is @OpenAI's biggest missed opportunity imo.
I went on a walk with ChatGPT Live and asked it to pull up my Google Doc. It said it couldn’t.
I then manually triggered the Documents plugin to
one thing i think people dont appreciate enough about @poolsideai is their unusual degree of openness — not only have they shipped an excellent Small model that somehow beat @thinkymachines at coding, but most people (like @eliebakouch) have been shouting out their excellent papers, but also they're among a rare few to actually expose their full eval dataset as well - beautifully published, with across 6 public benchmarks with 4 runs each and hundreds of turns per run. you can satisfy for yourself if they rewardhack. brilliant.
Latent.Space: Poolside’s Race to AGI: Model Factory, Laguna S, Open Models, & Beyond MCP https://latent.space/p/poolside
@poolsideai co-founder and CEO @eisokant explains why coding is the path to AGI, how Poolside’s Model Factory runs 10,000–20,000 experiments a month and turns new models around in
Marina Mogilko
The worry I keep hearing from creators at events: everyone is suddenly a creator, and the space feels oversaturated. I put it to @petergyang on my podcast, Silicon Valley Girl. He sees a lot of creators making slop, the 10-posts-an-hour kind, and it goes viral. He never wants to be the one making it, even when it pays. It's why he says finding your principles and values matters. And he thinks, maybe, people will get tired of it and come asking for human craft and human touch. His newsletter is called Behind the Craft. The full interview is on my channel.
Peter Yang: I’m sick of reading AI slop, so today I’m open-sourcing my /no-ai-slop skill that removes 20+ slop patterns from any piece of writing.
📌 Get the free skill here: https://github.com/petergyang/no-ai-slop
If you find it useful, please consider starring the repo so more people can find it.
Why I
Garry's List
A software engineer making $185K can't even afford to rent a one-bedroom in San Francisco.
A good salary used to buy you a home. Now it barely covers rent. For decades, we've made it nearly impossible to build, so even top earners are priced out.
The new fault line in American politics isn't rich vs. poor. It's people who own, and people who earn.
Tony Xu
Great to see @andyfang and @stanleytang talk about some of our latest experiments, some of which are now 8 years old, but still just getting started. If you want to learn, feel, and build at the intersection of autonomy, applied AI, and real world operations with speed and scale, come join us @DoorDash!
No Priors: .@DoorDash is betting on robotics, drones, and AI. That doesn't mean fewer Dashers. It means a much bigger delivery network.
@andyfang @stanleytang @saranormous
voice makes you feel just how unnatural it is to type
OpenAI: ChatGPT Voice is now in the desktop app.
Control your computer and direct multiple agents running in ChatGPT Work or Codex, using just your voice.
It's powered by GPT-Live, so it can speak, listen, and coordinate work in the app at the same time.
Rolling out globally today
I cannot find this feature anywhere in the ChatGPT desktop app, can you?
OpenAI: ChatGPT Voice is now in the desktop app.
Control your computer and direct multiple agents running in ChatGPT Work or Codex, using just your voice.
It's powered by GPT-Live, so it can speak, listen, and coordinate work in the app at the same time.
Rolling out globally today
