Fidji Simo

Companies go through phases of exploration and phases of refocus; both are critical. But when new bets start to work, like we're seeing now with Codex, it's very important to double down on them and avoid distractions. Really glad we're seizing this moment.

@Simonkhalaf Thanks @Simonkhalaf! But a lot of people contributed to News Feed :) So much of my career has been guided by the belief that we are all born creators, and the passion of putting tools of creation into people’s hands. And we are now witnessing the golden age of creation.

@fidjissimo @amyrochlin @CODA_research Just finished reanalyzing our Long Covid patient dataset with GPT-5.4 Pro. It gave me new insights we had missed! AI will be the most important facilitator in treating these dreadful & highly complex chronic illnesses. The biggest hurdle now is clinical testing the treatments.

@DeryaTR_ @amyrochlin @CODA_research That’s so amazing to read. Thanks for all you’re doing @DeryaTR_ . Hopefully AI can also accelerate clinical trials…

So grateful to @amyrochlin and the @CODA_research team for all they’re doing to advance research for the tens of millions of patients dealing with neuroimmune conditions. Every day I open my feed to amazing news of medical breakthroughs or other longevity magic, yet they stand in sharp contrast to all the tweets I see from patients who are bedbound, fighting for their lives, with not a single cure in sight. These debilitating diseases are sidelining millions and as a society, we are failing these people. We need to do better.

This news came out a little earlier than we planned; we're excited to be building a deployment arm and will share more details soon. Companies have a ton of urgency to deploy AI in their organizations and we’re sprinting to meet that demand. More than 1 million businesses run on OpenAI products. Codex is now at 2M+ weekly active users, up nearly 4x since the start of the year. API usage jumped 20% in the week after GPT-5.4 launched. And Frontier, which launched last month to help enterprises build, deploy, and manage AI coworkers that can do real work, has way more demand than we can handle. That's why we launched Frontier Alliances so we leverage our ecosystem of partners to scale. And that is also why we are launching a dedicated deployment arm tasked with embedding Forward Deployed Engineers deeply inside of enterprises.  This project has been in the works with our investor and alliance partners since last December, and we are grateful for them and their partnership. We’re still early, but the speed of adoption is a clear signal of where this is headed. We're excited to not just be building these technologies but also building many ways for companies to deploy them and get impact. reuters.com/business/opena…

I solved a problem with GPT that my doctor could not solve for YEARS. I was getting constantly sick to my stomach. Saw her a dozen times during that time. Saw specialists. Had an endoscopy (fun). Tried all kinds of different medicines. Different diets. Blood tests. Nothing worked. Eventually I figured it out with Reddit and GPT. It was my cholesterol medication. A rare side effect. Told my doctor. We changed to an alternative, GPT suggested might work and I double checked on WebMD. Problem went away in a week and never came back. Maybe you don't want people to have this power. Then you are my enemy and the enemy of freedom. I want to have this power at my finger tips and if you want to take it away from me I will fight you tooth and nail to keep it. People who want to take this power from you are protectionists, protecting existing guilds and incumbents, or they think you're too stupid and that you're a baby who has to be protected from yourself. Either way, they deserve zero respect and no quarter. They need to be beaten back at all levels of society because they are destroying it.

this is actually insane > be tech guy in australia > adopt cancer riddled rescue dog, months to live > not_going_to_give_you_up.mp4 > pay $3,000 to sequence her tumor DNA > feed it to ChatGPT and AlphaFold > zero background in biology > identify mutated proteins, match them to drug targets > design a custom mRNA cancer vaccine from scratch > genomics professor is “gobsmacked” that some puppy lover did this on his own > need ethics approval to administer it > red tape takes longer than designing the vaccine > 3 months, finally approved > drive 10 hours to get rosie her first injection > tumor halves > coat gets glossy again > dog is alive and happy > professor: “if we can do this for a dog, why aren’t we rolling this out to humans?” one man with a chatbot, and $3,000 just outperformed the entire pharmaceutical discovery pipeline. we are going to cure so many diseases. I dont think people realize how good things are going to get

ChatGPT's retention is in a league of its own. 71% of users still active at Month 10... nearly 2x the next closest AI app. Most consumer apps would kill for these numbers. Full analysis in Part 2 of our State of Consumer AI series 👇 x.com/apoorv03/statu…

If you want AI Code Review, but don't want to pay $25 per review (not a typo), check out Codex Review! It leverages frontier Codex models, finds complex issues, and 100% usage based. Most runs should cost ~$1 or less developers.openai.com/codex/integrat…

Chatgpt Vs Claude in Excel TLDR: Gave ChatGPT and Claude the same credit risk modeling task in Excel. ChatGPT followed proper methodology, did variable clustering, train/test split, Excel native model, AUC 0.627 on test data. Claude skipped clustering, inflated IV with sparse categories, ran the model in JavaScript not Excel, reported 0.706 AUC on training data with no split. Chagpt 5.4 nailed a real world Banking and Finance modelling case study. I gave both ChatGPT and Claude the same banking credit risk dataset with the same prompt: bin the variables, calculate Information Value, do variable clustering, select features, build a model in Excel, generate ROC. Basically an end to end scorecard development workflow inside Excel. ChatGPT followed the brief properly. Quantile-based WoE binning, IV ranking across all 33 variables, correlation-based variable clustering at 0.75 threshold to remove redundant features, picked 6 representative variables, built a 2-variable Excel-native decision tree with proper train/test split via ID mod 10, scored leaf-level bad rates, ROC on held-out test data. AUC: 0.627. Claude looked more impressive on the surface. 28 sheets, individual binning for every variable, logistic regression with 27 features, gradient descent at 500 epochs, full coefficient table with importance bars, confusion matrix, precision/recall/F1. AUC: 0.7066. Sounds like it won right? So, The logistic regression was computed in JavaScript via execute_office_js and results were pasted as static values. That's not Excel-native, against the task. The 0.706 AUC? On the full training set, no train/test split. That number is meaningless. Variable clustering? Skipped entirely even though the prompt explicitly asked for it. It "selected" 31 out of 33 features which is barely feature selection. The entire feature ranking was built on a methodological error. I've done this exact dataset myself with deeper feature engineering and ensemble technique hit 0.70 AUC. The fact that ChatGPT got to 0.627 with less feature enigineering and naive modelling is genuinely impressive. Binning and IV based feature selection is one of the most important techniques in credit risk modeling and it nailed the workflow. One thing claude did well is the workbook is visually impressive. One area ChatGPT can improve, the visuals and formatting could be sharper, the workbook is functional but not polished.

$200/mo ChatGPT Pro + gpt-5.4 xhigh + Codex Mac App is the most asymmetric upside I've ever seen in a tool chain. A ridiculous cheat code for founders.

5.4 is faster and better at professional work -- with big improvements in spreadsheet, doc, and slide creation. In Codex and the API, it's our first general purpose model with native SOTA computer use capabilities, which is going to enable so much more agentic work.

Codex app on Windows!

We believe it is incredibly important for the government to have access to AI under the right conditions; I hope we can all agree that the safety of Americans should trump any revenue objective so we really should not need any big incentive to want to help keep our country safe. It’s an extremely cynical take to believe we would only do the right thing here if it was worth billions to us; it’s pretty apparent from this conversation that it’s costing us a lot instead. Despite the cost, it remains the right thing to have done. As for the difference between our deal and Anthropic’s, the main difference is that we came up with a way to enforce our redlines through technical implementations (like relying on our own safety stack, which can refuse to take certain actions and embeds our red lines into code, and deploying our engineers at DoW, which the admin was comfortable with) and not just through usage policies. We actually believe that these technical implementations are more robust than contractual ones, but apparently weren’t explored by Anthropic (credit to @sama for coming up with a solution that was both more robust but more acceptable to the govt). Now that these are in place, we hope the government extends that same approach to all AI labs and that these AI labs accept it. We also believe that this approach keeps things that should be controlled by AI labs (like what we allow our models to do) in our camp, while keeping things that are definitely the prerogative of the administration (national security strategy and which actions to take to protect the country) in the hands of our democratically elected government and not in the hands of the unelected CEO of a private company.

Do you really expect people to believe you got involved in a deal with this much risk for an “inconsequential” reward? With all due respect, give me a break. Getting involved here implicitly carries risk of being labeled a supply chain risk should things turn sour…not to mention the tremendous negative PR realized immediately. This was either the most incompetent decision I’ve ever seen—taking inordinate risk for minimal gain, or you are woefully mischaracterizing how much you stand to benefit financially from this. And furthermore, your blog post does not assuage any of my concerns. You fail to explain how you were able to fill the role Anthropic did, and do the things they were so concerned about doing that it warranted blowing up their contract…all without any of the same concerns. The issue is not developing some structure of “guardrails” out of the air—referencing the same legal constraints that bound and were apparently insufficient for Anthropic’s purposes—it’s about explaining in relative terms how and why you are able to do something your direct competitor refused to do out of concern for safety and ostensibly their customers’ privacy…without exposing your own customers to the same risk. Until you can explain that concisely, I personally will not feel comfortable using your products. At least in ways that involve trust and my data.

I have been a loyal @OpenAI guy from day 1. I’ve had an account since they first released and used it daily since I think GPT 3.5. Based on @sama’s move tonight to accept the DoD terms that Anthropic would not, I am removing all traces of Codex and OpenAI software from all devices. I absolutely hate to do this and it’s going to really compromise my ability to be productive in the short term, but every possible alarm bell is going off at full blast in my head right now. This has the potential to be the most egregious incursion on privacy the world may have ever seen. I think Sam just saw too many dollar signs and didn’t realize how obvious this move would make the optics but it’s clear to me now OpenAI is compromised. Not sure what the next step will be from here, but the trust is totally gone. Very disappointing.

@uday_devops @sama it's a few million $, completely inconsequential compared to our $20B+ in revenue, and definitely not worth the cost of a PR blowup. We’re doing it because it's the right thing to do for the country, at great cost to ourselves, not because of revenue impact

Personal view. On the DoW deal, time and time again I witness how OpenAI operates thoughtfully and diplomatically when it comes to raising the bar on safety. I also believe that figuring out how to deploy powerful new technology in the pursuit of national security (not just the USA) is important. The company deeply understands what it can and cannot reasonably have control over and focuses instead of setting the right guardrails to ensure that deployment is aligned with what is considered acceptable and safe. I am proud of the level of transparency provided in the blog post, and hope that other companies, including Anthropic, were as transparent on prior and future deals they make. openai.com/index/our-agre…

Our agreement with the Department of War upholds our redlines: - No use of OpenAI technology for mass domestic surveillance. - No use of OpenAI technology to direct autonomous weapons systems. - No use of OpenAI technology for high-stakes automated decisions (e.g. systems such as “social credit”).