第52期 | Enterprises power agentic workflows in Cloudflare Agent...
今日摘要
OpenAI Blog：OpenAI launches Codex Labs, partners with with Accenture, PwC, Infosys, and others to help enterprises deploy and scale Codex acro…
OpenAI Blog：Hyatt deploys ChatGPT Enterprise across its global workforce, using GPT-5.4 and Codex to improve productivity, operations, and gue…
OpenAI Blog：The updated Codex app for macOS and Windows adds computer use, in-app browsing, image generation, memory, and plugins to accelerat…
OpenAI Blog：OpenAI introduces GPT-Rosalind, a frontier reasoning model built to accelerate drug discovery, genomics analysis, protein reasonin…
OpenAI Blog：Leading security firms and enterprises join OpenAI’s Trusted Access for Cyber, using GPT-5.4-Cyber and $10M in API grants to stren…
总结 + 观点：OpenAI updates the Agents SDK with native sandbo…｜中文观点：The next evolution of the Agents SDK 的价值在于它是否…
总结 + 观点：OpenAI expands its Trusted Access for Cyber prog…｜中文观点：从 Trusted access for the next era of cyber de…
总结 + 观点：Cloudflare brings OpenAI’s GPT-5.4 and Codex to…｜中文观点：从 Enterprises power agentic workflows in Clou…
总结 + 观点：Learn how sales teams use ChatGPT to research ac…｜中文观点：对 ChatGPT for sales teams，更该看它能不能改善多步骤协作、记忆管理…
总结 + 观点：Learn how to use ChatGPT, start your first conve…｜中文观点：围绕 Getting started with ChatGPT，真正重要的是它会不会影响团…
Scaling Codex to enterprises worldwide
标签：#ai_engineering_blogs #core
作者：
原文：OpenAI launches Codex Labs, partners with with Accenture, PwC, Infosys, and others to help enterprises deploy and scale Codex across the software development lifecycle, and hits 4M Codex WAU.
链接：https://openai.com/index/scaling-codex-to-enterprises-worldwide
OpenAI helps Hyatt advance AI among colleagues
标签：#ai_engineering_blogs #core
作者：
原文：Hyatt deploys ChatGPT Enterprise across its global workforce, using GPT-5.4 and Codex to improve productivity, operations, and guest experiences.
链接：https://openai.com/index/hyatt-advances-ai-with-chatgpt-enterprise
Codex for (almost) everything
标签：#ai_engineering_blogs #core
作者：
原文：The updated Codex app for macOS and Windows adds computer use, in-app browsing, image generation, memory, and plugins to accelerate developer workflows.
Introducing GPT-Rosalind for life sciences research
标签：#ai_engineering_blogs #core
作者：
原文：OpenAI introduces GPT-Rosalind, a frontier reasoning model built to accelerate drug discovery, genomics analysis, protein reasoning, and scientific research workflows.
Accelerating the cyber defense ecosystem that protects us all
标签：#ai_engineering_blogs #core
作者：
原文：Leading security firms and enterprises join OpenAI’s Trusted Access for Cyber, using GPT-5.4-Cyber and $10M in API grants to strengthen global cyber defense.
链接：https://openai.com/index/accelerating-cyber-defense-ecosystem
The next evolution of the Agents SDK
标签：#ai_engineering_blogs #core
作者：
原文：OpenAI updates the Agents SDK with native sandbox execution and a model-native harness, helping developers build secure, long-running agents across files and tools.
链接：https://openai.com/index/the-next-evolution-of-the-agents-sdk
Trusted access for the next era of cyber defense
标签：#ai_engineering_blogs #core
作者：
原文：OpenAI expands its Trusted Access for Cyber program, introducing GPT-5.4-Cyber to vetted defenders and strengthening safeguards as AI cybersecurity capabilities advance.
链接：https://openai.com/index/scaling-trusted-access-for-cyber-defense
Enterprises power agentic workflows in Cloudflare Agent Cloud with OpenAI
标签：#ai_engineering_blogs #core
作者：
原文：Cloudflare brings OpenAI’s GPT-5.4 and Codex to Agent Cloud, enabling enterprises to build, deploy, and scale AI agents for real-world tasks with speed and security.
ChatGPT for sales teams
标签：#ai_engineering_blogs #core
作者：
原文：Learn how sales teams use ChatGPT to research accounts, personalize outreach, manage deals, and improve pipeline and conversion.
Getting started with ChatGPT
标签：#ai_engineering_blogs #core
作者：
原文：Learn how to use ChatGPT, start your first conversation, and discover simple ways to write, brainstorm, and solve problems with AI.
Someone recently suggested to me that the reason OpenClaw moment was so big is because it's the first time a large group of non-technical pe...
标签：#x_profiles #extended
作者：
原文：Someone recently suggested to me that the reason OpenClaw moment was so big is because it's the first time a large group of non-technical people (who otherwise only knew AI as synonymous with ChatGPT as a website) experienced the latest agentic models.
Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use.
标签：#x_profiles #extended
作者：
原文：Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are. staysaasy (@staysaasy) The degree to which you are awed by AI is perfectly correlated with how much you use AI to code. https://nitter.net/staysaasy/status/2042063369432183238#m
karpathy/nanochat
标签：#github_orgs #extended
作者：
原文：Karpathy 新发布的最小 ChatGPT 复现项目，训练到推理的完整栈只有几千行可读代码，目标是把“百美元跑一个 ChatGPT”压到个人可动手的范围。
karpathy/minGPT
标签：#github_orgs #extended
作者：
原文：Karpathy 早期的教学级 GPT 实现，代码短到可以一口气读完，长期用作理解 Transformer 训练与推理最短路径的入口。
anthropics/original_performance_takehome
标签：#github_orgs #extended
作者：
原文：Anthropic 公开其内部工程师 take-home 面试题，可作为理解他们工程品味和评估标准的一手材料。
链接：https://github.com/anthropics/original_performance_takehome
Drunk Post: Things I've Learned as a Senior Engineer
标签：#research_community #core
作者：
原文：Article URL: https://luminousmen.substack.com/p/drunk-post-things-ive-learned-as Comments URL: https://news.ycombinator.com/item?id=47856535 Points: 39 Comments: 11
链接：https://luminousmen.substack.com/p/drunk-post-things-ive-learned-as
SpaceX says it has agreement to acquire Cursor for $60B
标签：#research_community #core
作者：
原文：https://www.reuters.com/technology/spacex-says-it-has-option... https://www.nytimes.com/2026/04/21/business/spacex-cursor-de... https://archive.ph/c2Tac https://www.bloomberg.com/news/articles/2026-04-21/spacex-sa... Comments URL: https://news.ycombinator.com/item?id=47855293 Points: 350 Comments: 470
Claude Code to be removed from Anthropic's Pro plan?
标签：#research_community #core
作者：
原文：https://x.com/TheAmolAvasare/status/2046725498592722972 https://xcancel.com/TheAmolAvasare/status/204672549859272297... Comments URL: https://news.ycombinator.com/item?id=47854477 Points: 388 Comments: 408
Where's the raccoon with the ham radio? (ChatGPT Images 2.0)
标签：#ai_engineering_blogs #core
作者：
原文：OpenAI released ChatGPT Images 2.0 today their latest image generation model. On the livestream Sam Altman said that the leap from gpt-image-1 to gpt-image-2 was equivalent to jumping from GPT-3 to GPT-5. Here's how I put it to the test. My prompt: Do a where's Waldo style image but it's where is the raccoon holding a ham radio gpt-image-1 First as a baseline here's what I got from the older gpt-image-1 using ChatGPT directly: I wasn't able to spot the raccoon - I quickly realized that testing image generation models on Where's Waldo style images (Where's Wally in the UK) can be pretty frustrating! I tried getting Claude Opus 4.7 with its new higher resolution inputs to solve it but it was convinced there was a raccoon it couldn't find thanks to the instruction card at the top left of the image: Yes there's at least one raccoon in the picture, but it's very well hidden In my careful sweep through zoomed-in sections, honestly, I couldn't definitively spot a raccoon holding a ham radio. Nano Banana 2 and Pro Next I tried Google's Nano Banana 2, via Gemini That one was pretty obvious, the raccoon is in the "Amateur Radio Club" booth in the center of the image! Claude said: Honestly, this one wasn't really hiding he's the star of the booth. Feels like the illustrator took pity on us after that last impossible scene. The little "W6HAM" callsign pun on the booth sign is a nice touch too. I also tried Nano Banana Pro in AI Studio and got this, by far the worst result from any model. Not sure what went wrong here! gpt-image-2 With the baseline established, let's try out the new model. I used an updated version of my openai_image.py script, which is a thin wrapper around the OpenAI Python client library. Their client library hasn't yet been updated to include gpt-image-2 but thankfully it doesn't validate the model ID so you can use it anyway. Here's how I ran that: OPENAI_API_KEY= llm keys get openai uv run https://tools.simonwillison.net/python/openai_image.py -m gpt-image-2 Do a where's Waldo style image but it's where is the raccoon holding a ham radio Here's what I got back. I don't think there's a raccoon in there - I couldn't spot one, and neither could Claude. The OpenAI image generation cookbook has been updated with notes on gpt-image-2 including the outputQuality setting and available sizes. I tried setting outputQuality to high and the dimensions to 3840x2160 - I believe that's the maximum - and got this - a 17MB PNG which I converted to a 5MB WEBP: OPENAI_API_KEY= llm keys get openai uv run https://raw.githubusercontent.com/simonw/tools/refs/heads/main/python/openai_image.py -m gpt-image-2 Do a where's Waldo style image but it's where is the raccoon holding a ham radio --quality high --size 3840x2160 That's pretty great! There's a raccoon with a ham radio in there (bottom left, quite easy to spot). The image used 13,342 output tokens, which are charged at $30/million so a total cost of around 40 cents Takeaways I think this new ChatGPT image generation model takes the crown from Gemini, at least for the moment. Where's Waldo style images are an infuriating and somewhat foolish way to test these models, but they do help illustrate how good they are getting at complex illustrations combining both text and details. Update: asking models to solve this is risky rizaco on Hacker News asked ChatGPT to draw a red circle around the raccoon in one of the images in which I had failed to find one. Here's an animated mix of their result and the original image: Looks like we definitely can't trust these models to usefully solve their own puzzles! Tags: ai openai generative-ai chatgpt llms text-to-image llm-release nano-banana
链接：https://simonwillison.net/2026/Apr/21/gpt-image-2/#atom-everything
Zindex Diagram Infrastructure for Agents
标签：#research_community #core
作者：
原文：Article URL: https://zindex.ai/ Comments URL: https://news.ycombinator.com/item?id=47854116 Points: 41 Comments: 15