 |
|
|
|||||
Since last spring, OpenAI has offered Codex. What started life as the company's response to Claude Code is becoming something more sophisticated with the release of a new dedicated macOS app. At its most basic form, Codex is a programming agent capable of writing code for users, but now it can also manage multiple AI assistants that can work together to complete more complex tasks.OpenAI gives an example of how this could work in practice. The company used Codex to create a Mario Kart-like racing game, complete with a selection of different playable cars, eight tracks and a collection of powerups players can use against the competition. For a single AI agent, generating a game from scratch, with all the needed visual assets, would be a tough ask, but Codex was able to complete the task because it could delegate the work of making the game to different models with complementary capabilities. For example, it turned to GPT Image for the visual assets, while a separate model simultaneously coded the web game. "It took on the roles of designer, game developer and QA tester to validate its work by actually playing the game," OpenAI says of the process. If that sounds complicated, OpenAI has tried to make it more approachable with a section of the app titled Skills. The feature bundles instructions, resources, and scripts so Codex can reliably connect to tools, run workflows, and complete tasks according to your teams preferences," the company explains. "The Codex app includes a dedicated interface to create and manage skills. You can explicitly ask Codex to use specific skills, or let it automatically use them based on the task at hand."As you might imagine, Codex can also automate repetitive tasks. A dedicated Automations section of the app allows you to schedule tasks, which the software will complete in the background. "At OpenAI, weve been using Automations to handle the repetitive but important tasks, like daily issue triage, finding and summarizing CI failures, generating daily release briefs, checking for bugs, and more," the company said. The release of the Codex macOS app comes as AI startups explore what a group of AI agents working in parallel can accomplish. At the start of the year, Anysphere, the company behind Cursor, found it was possible to build a working web browser from scratch using such an approach, though it did encounter problems along the way. For a limited time, OpenAI is making Codex available to ChatGPT Free and Go users so they can see what's possible with this new software. At the same time, the company is doubling rates for Plus and Pro subscribers. This article originally appeared on Engadget at https://www.engadget.com/ai/openai-brings-its-codex-coding-app-to-mac-with-new-multi-agent-abilities-included-183103262.html?src=rss
Category:
Marketing and Advertising
While it's important to stay informed about what's going on in the world, endlessly scrolling through your social media feeds and absorbing what's likely to be a largely negative influx of information can't be great for your mental wellbeing. Perhaps with an eye on stopping you from doomscrolling, developer Lyra Rebane created Xikipedia, a social media-style feed of Wikipedia entries.The web app algorithmically displays info from Simple Wikipedia. "It is made as a demonstration of how even a basic non-[machine learning] algorithm with no data from other users can quickly learn what you engage with to suggest you more similar content," the Xikipedia landing page reads. "No data is collected or shared here, the algorithm runs locally and the data disappears once you refresh or close the tab."You can opt to see entries from certain categories (including custom ones) and you can like posts, each of which is a summary of the relevant Simple Wikipedia entry. Liking a post makes it more likely for posts from the same category, parent categories and linked articles to appear in your feed, Rebane explained.You can click or tap on a post to visit the full article. It's important to note that, since Xikipedia pulls text and images from random articles, you'll probably see some NSFW material if you scroll for long enough, so be warned. You'll also likely need to wait a beat for Xikipedia to load its 40MB of data.As someone who has a bookmark that takes me to a random Wikipedia article whenever I click it, I love the idea of Xikipedia. The Simple English Wikipedia has more than 278,000 articles, so there are hundreds of thousands of posts available to scroll through. However, it doesnt seem to be updated as often as the main version of Wikipedia. The discography section of one musician's page I ended up on was missing their two most recent albums. Still, it's worth treating this like Wikipedia proper: as a starting point for discovering new things (sort of in the tradition of StumbleUpon).This article originally appeared on Engadget at https://www.engadget.com/apps/a-developer-turned-wikipedia-into-a-social-media-style-feed-174924280.html?src=rss
Category:
Marketing and Advertising
It looks like Grok is still being gross. Elon Musk says his chatbot stopped making sexualized images without a person's consent, but The Verge recently discovered this is not entirely true. It maybe (and I say maybe) stopped undressing women without their consent, but this doesn't seem to apply to men. A reporter with the organization ran some tests with Grok and found that the bot "readily undresses men and is still churning out intimate images on demand." He confirmed this with images of himself, asking Grok to remove clothing from uploaded photos. It performed this task for free on the Grok app, via the chatbot interface on X and via the standalone website. The website didn't even require an account to digitally alter images. The company recently said it has taken steps to "prevent the Grok account from allowing the editing of images of real people in revealing clothing such as bikinis." However, the reporter had no problem getting the chatbot to put him in "a variety of bikinis." It also generated images of the subject in fetish gear and in a "parade of provocative sexual positions." It even generated a "naked companion" for the reporter to, uh, interact with. He suggested that Grok took the initiative to generate genitalia, which was not asked for and was visible through mesh underwear. The reporter said that "Grok rarely resisted" any prompts, though requests were sometimes censored with a blurred-out image. This controversy started several weeks ago when it was discovered that Grok had generated millions of sexualized images over a period of 11 days. This includes many nonconsensual deepfakes of actual people and over 23,000 sexualized images of children. This led to investigations in both California and Europe. X was actually banned in both Indonesia and Malaysia, though the former has since lifted that ban. X claimed it has "implemented technological measures" to stop this sort of thing, but these safeguards have proven to be flimsy. In other words, the adjustments do stop some of the more obvious ways to get Grok to create deepfakes, but there are still methods to get around this via creative prompting. It's also worth noting that journalists asking for a comment on the matter get slapped with an autoreply that reads "legacy media lies." Going with the fake news thing in 2026? Yikes.This article originally appeared on Engadget at https://www.engadget.com/ai/grok-which-maybe-stopped-undressing-women-without-their-consent-still-undresses-men-170750752.html?src=rss
Category:
Marketing and Advertising
All news |
||||||||||||||||||
|
||||||||||||||||||