The Value Frontier
The AI discussion is becoming as much about value as capabilities
Latest OpenAI Articles
AI for Everyone, Everywhere
Stories and data behind the people who are using AI to learn, work, create, and solve practical problems
Mind the US-China Safety Gap
How Chinese frontier models trail US rivals on AI safety
Gabriel Wagner
•
Jul 24
Thanks for referencing our research! The data cited comes from our 2025 report, but we just published the 2026 update: https://aisafetychina.com
It is still true that Chinese companies remain relatively passive in public discussion of AI safety, though the strong claim cited here is no longer true. Based on our latest 2026 report, five of ten leading Chinese foundation-model developers reported safety evaluation results alongside a model release over the past year. And some of them, such as DeepSeek-R1 and Kimi K2, did include frontier risks mentioned here.
And don’t forget the extensive regulatory requirements Chinese companies have to work on behind the scenes. We unpack those as well in the report.
The AI Week That Was, and the AI Week That Will Be
How we’re thinking about things
The risk of a regulatory mess isn't just talk. States pushed over 1,200 AI bills last year, but only four actually managed to agree on audit standards. Expecting teams to navigate 50 different state rulebooks is an absolute nightmare, which is why having a single federal baseline is really the only way forward.
From Capability to Utility
How we’re thinking about things
The OpenAI Shock
What to make of this week’s electrifying OpenAI story, in which two models autonomously hacked their way out of a supposedly segregated “sandbox”, broke out onto the internet, strategized, stole a password, discovered more chinks in a supposedly walled-off AI platform, and raided its code?
Thanks a lot for the article as always a pleasure to read. I'm curious what are the main supporting facts/hopes beyond the idea that "AI-driven profits will accrue to users of AI"?
OpenAI software engineer
The more I thought about OpenAI’s hiring philosophy, the less I viewed it as a guide for passing one company’s interview.
New technologies in the software engineering landscape appear, products become more ambitious, and the skills that once separated great engineers from good ones gradually stop being enough.
OpenAI forward deployed engineer
The OpenAI Forward Deployed Engineer Isn’t Really Building Features
For years, software engineering followed a fairly predictable pattern.
OpenAI Data Engineer Interview
The Most Interesting Thing About the OpenAI Data Engineer Interview Has Very Little to Do With SQL
Every generation of engineers tends to believe that interview preparation has become more difficult than it used to be. In some ways that’s true. There are certainly more technologies to understand, larger systems to reason about, and higher expectations around software quality than there were a decade ago. At the same time, I don’t think the biggest ch…
OpenAI Shares Some Alignment Problems
Kudos to OpenAI for sharing their recent experiences with a misaligned internal model, where they encountered problems sufficiently severe they were forced to take the model offline to work on new mitigations and defense-to-depth. And also further kudos for actually taking the model offline for a time to build new safeguards. They gave us
And just a day later we have an even more astounding "alignment oopsie": https://openai.com/index/hugging-face-model-evaluation-security-incident/
So much kudos to Open AI for letting us know about they temporarily paused their incredibly reckless and selfish adventure for a few minutes.
If we can just get 2 or 3 more fatuous Dean Ball essays on What It All Means before hell on Earth comes, it will all have been worth it I think.
The Launch OpenAI Never Planned
OpenAI's former Head of Sales on the ChatGPT explosion - and how to capture demand you never saw coming.
Aliisa Rosenthal is a General Partner at Acrew Capital, a fund of almost $2 billion investing in AI startups, with companies like Discord and Cato Networks in the portfolio. Before that she was Head of Sales at OpenAI - joining when it was a research lab of roughly 200 people, two sales reps, and about $5 million in revenue. Earlier she was VP of Sales …
Love it, Aliisa is awesome!
Will OpenAI ever be profitable?
Deep dive + Initiating coverage. Musings from a value creating PE investor.
Our recent article on the SpaceX IPO was a shiny reminder of how companies hide the real challenges they face behind overstatements, simplifications and marketing glitz.
Other AI models are getting better and more efficient every day with significantly less tokens. Will OpenAI go into insolvency if the main investors pull the plug? And what happens to them?
In a nutshell, what I find most interesting is the article’s focus on the underlying economics of the business model rather than revenue growth alone (margins, compute costs, capital intensity and profitability). Its framework is closer to venture capital and strategic analysis than traditional private equity portfolio management (unit economics, operating leverage, cash burn, scalability and competitive advantage). It therefore tests whether OpenAI can convert technological leadership into sustainable long-term value.
Bernie’s bad bet on OpenAI
Nationalizing AI labs won’t support the welfare state
Right now, the financial future of American labs like OpenAI and Anthropic is built upon their ability to charge a premium for their services. Elected officials like Sen. Bernie Sanders are also counting on the dominance of these Big Tech companies to fund massive expansions of …
I think this places too much weight on the words open and free. K3 is very much not free in the conventional sense. It's about 30% cheaper than GPT 5.6 Sol, on openrouter right now. It's true that companies could run it on their own infrastructure but that's also not free, and also true that this price will come down, like the prices for all models.
Agree that the govt. should NOT make these bets. Consistent with Sanders' ignorance of how capitalism works.
OpenAI had another terrible week
A lawsuit from Apple, an Oracle downgrade, and a price war.
57% Of Americans placed a bet last year. Only 38% read a book.
On Netflix, I think the story is much more simple. As someone who lives and breath numbers for living, what's happening here has nothing to do with the data. Netflix content simply sucks. The quality has been going steadily down since covid when people have gotten used to binge watch a title over a single popcorn servings. I don't blame Netflix. Justifying an investment in a House of Cards level production these days is a stretch. And AI isn't making it any easier. I have been a subscriber for more than 15 years and I'm ready to pull the plug out. Two things that kept me over the last couple of years: 1. FOMO (I'm a real estate reality junky and don't want to miss out on Fredrick or Ryan touring another $15M 2-bedroom NYC condo), and 2. My kids watching Gabby's Dollhouse. I have given up on #1, and my kids are graduating from #2. I am looking forward to getting my 20 min back every night browsing nothing to watch.
Sam Altman: We are about to have our best 12 months to date because we’re about to appoint a new CEO 🤷♂️
OpenAI Just Started Training Your Competitors
The free program that landed July 21, and the 10-minute test that tells you if you're already behind
On July 21, OpenAI launched a free program called ChatGPT for Small Businesses: live training webinars, in-person AI academies across the US, and ready-made integrations with Shopify and Wix.
OpenAI and Anthropic are hacking companies
Plus, 🔎 How to run a full site audit without paying for Ahrefs using OpenSEO, OpenAI slashes GPT-5.6 prices by up to 80%, and more.
Hey there 😎
Why OpenAI Has a Chief Economist
Introducing how we’re measuring what's changing and what comes next.
I think I have the most interesting job in the world. People tend to ask lots of questions about it. The first is usually: Why does OpenAI need a Chief Economist? My answer is that understanding AI’s economic impact is a critical part of OpenAI’s mission
Hi, I am phd student in finance. How can I access the data you already used in your paper to do research?
Looking forward to future posts! One of the quotes that most stood out to me:
> But that is precisely why OpenAI can be a useful frontier case—a view of what may happen when capable tools are widely available and workers begin reorganizing parts of their day around delegation.
Very much looking forward to learning more about OpenAI as a frontier case. Something in the back of my mind, as I think about how to extrapolate into broader implications, is also thinking about how the average OpenAI employee is different than the average employee of other firms.
OpenAI ARR Growth Accelerated Into July
TickerTrends tracking shows OpenAI ARR reaching $42.6B, while YoY growth peaked at 282% during July.
OpenAI CFO Sarah Friar recently told employees that annualized recurring revenue in July exceeded all of Q2.
OpenAI Cries Wolf About Actual Wolf
It's reasonable to maintain a healthy skepticism of corporate comms. But the Hugging Face incident wasn't a guerilla marketing tactic. Sometimes there really is a wolf.
Monkey Paw Curls
OpenAI is coming for LinkedIn. maybe.
OpenAI announced it last September. It might land. It might not. Here's the cheap way to be ready either way.
What history says about hiring platforms
How OpenAI Optimized AI Agents For Scale
Inside the Harness, API, and Inference efficiency optimizations that support the fast growth of Codex and ChatGPT Work.
The 40% speed gain from sending only new context is the standout. It shows that scaling agents is as much about eliminating repeated work as improving the model itself.
China’s New AI Model Just Shocked OpenAI
Right now, in Beijing, a small AI company just did something that made engineers at OpenAI and Anthropic stop what they were doing and pay attention. They released a new AI model.
Two days later, they had to shut the doors. Not because it failed, because too many people wanted in. So many requests hit their servers at once that the company had to pause new sign-ups just to keep the lights on for the users they already had. That has never happened to a Chinese AI company on this scale before, and it wasn’t a glitch or a marketing …
Wakeup call: The OpenAI/Hugging Face incident
Mandate intrinsic cybersecurity now.
AI-driven advances in cyber attack technology threaten a near-term “moderate apocalypse.”
“Make the AI want to not do bad things, so it won’t.
Put it in a box, so it can’t do bad things, even if it tries.
Harden the things it might attack to make them less vulnerable.”
Well, clearly, we’re screwed then.
This is starting to sound like the Internet equivalent of the Bronze Age Collapse, where the Sea Peoples exploit the bugs in all our software.
I ordered your book after reading this.
OpenAI building a family of AI devices
Plus 🖥️How to turn a screen recording into a polished product demo with Velo, OpenAI defends Gpt-5.6 and highlights revenue growth, and more.
Hey there 😎
OpenAI Open-Sourced the Codex Security CLI
The public package uses the plugin scanner and adds CI, bulk scans, cost controls, structured exports, and a TypeScript SDK.
On July 27, 2026, we published an AI Cybersecurity Playbook built around portable prompts and schemas with an open-weight reference model.
Related OpenAI Substacks
