🚀 Zaprep: Tus redes sociales al máximo. Empezar gratis — automatiza 1,000 DMs/mes y convierte el engagement en leads. con 1,000 DMs automatizados/mes.
Enviar tu herramienta
Últimas Noticias de IA
OpenAI updates its Agents SDK to help enterprises build safer, more capable agents
Agentic AI is the tech industry’s newest success story, and companies like OpenAI and Anthropic are racing to give enterprises the tools they need to create these automated little helpers. To that end,OpenAI has now updated its agents software development toolkit (SDK), introducing a number of new features designed to help businesses create their own agents that run on the backs of OpenAI’s models. The SDK’s new capabilities include a sandboxing ability, which allows the agents to operate in controlled computer environments. This is important because running agents in a totally unsupervised fashion can beriskydue to their occasionally unpredictable nature. With the sandbox integration, agents can work in a siloed capacity within a particular workspace, accessing files and code only for particular operations, while otherwise protecting the system’s overall integrity. Relatedly, the new version of the SDK also provides developers with an in-distribution harness for frontier models that will allow those agents to work with files and approved tools within a workspace, the company said. (In agent development, the “harness” is a term that refers to the other components of an agent besides the model that it’s running on. An in-distribution harness often allows companies to both deploy and test the agents running on frontier models, whichare considered to bethe most advanced, general-purpose models available.) “This launch, at its core, is about taking our existing agents SDK and making it so it’s compatible with all of these sandbox providers,” Karan Sharma, who works on OpenAI’s product team, told TechCrunch. The hope is that this, paired with the new harness capabilities, will allow users “to go build these long-horizon agents using our harness and with whatever infrastructure they have,” he said. Such “long-horizon” tasks are generally considered to be more complex and multi-step work. OpenAI said it will continue to expand the Agents SDK over time, but initially, the new harness and sandbox capabilities are launching first in Python, with TypeScript support planned for a later release. The company said it’s also working to bring more agent capabilities, like code mode and subagents, to both Python and TypeScript. The new Agents SDK capabilities are being offered to all customers via the API, and will use standard pricing.
View
This startup is betting tokenmaxxing will create the next compute giant
“Give me tokens. Just give me tokens. I want them fast. I want them cheap. I want them now.” That’s the mantra for developers building software on generative AI models, or at least what Parasail CEO Mike Henry hears. Parasail provides a cloud computing service to companies running AI models for inference, and Henry told TechCrunch it generates 500 billion tokens a day. How’s that for tokenmaxxing? Henry was an executive at Groq, the LLM-focused chipmaker, where he built the company’s cloud offering, an early recognition that developers building software on AI models would want cloud processing specialized to their needs. Now, aftercoming out of stealtha year ago, Parasail has raised a $32 million Series A to do that at scale. Henry has a background in physical chip design, but Parasail isn’t committed to owning its own chips. While some of its GPUs are its own, the company mainly rents processing time at 40 data centers in 15 countries around the globe, and buys more from liquidity markets, orchestrating that all behind the scenes to drive down the cost of inference requests. By allocating workloads cleverly and avoiding demand peaks, the company aims to compete with firms that own their own silicon and might be constrained by existing customer commitments and workloads. The company’s potential relies on the continued proliferation of open-source models and agents outside of frontier labs. Parasail’s executives and investors say this is driven by the growing cost and friction of using offerings from companies like Anthropic and OpenAI. Instead, a hybrid architecture is emerging, according to Andreas Stuhlmüller, the CEO of Elicit, a startup that has raised a $22 million Series A to develop a research assistant for scientific literature. His customers at top pharmaceutical companies use the LLM-based tool to review and analyze data from tens of thousands of scientific papers. “We’ve moved more towards open models because it’s pretty rough sending 100,000s of requests to an API endpoint,” Stuhlmüller told TechCrunch, especially now that the company is relying on agents to improve its offering, splitting up tasks and working more strategically over longer time horizons. Open models handle the initial screening to drive down the cost of the work, before a more capable frontier model provides a final answer. The proliferation of model queries, as agents become an increasingly common part of software development, is driving the investment in companies like Parasail that provide the infrastructure for cheap inference. Samir Kumar, a partner at Touring Capital who co-led this round, told TechCrunch he expects inference to be at least 20% of the cost of building software in the future. How much of that market could be Parasail’s? In the crowded cloud compute space, Henry argues that his firm’s focus on inference (no training allowed) and willingness to take on startup customers without long-term commitments sets his offering apart from larger cloud-computing companies focused on enterprise business, and even better-funded competitors in the cloud inference space, like Fireworks AI and Baseten. Of course, there’s a different kind of risk when all of your customers are seed and Series B startups in the unpredictable AI sector. Steve Jang, a partner at Kindred Ventures, the other co-leader in this fundraising, says the economics of deploying models will demand the kind of compute brokerage Parasail provides. And that’s before widespread use of models for content generation and robotics. “Everyone thought there was an AI bubble. There’s no AI bubble,” he told TechCrunch. “Inference demand is far outstripping supply.”
View
Gitar, a startup that uses agents to secure code, emerges from stealth with $9 million
With the advent of “vibe coding,” AI agents haveunleashed a deluge of codeonto companies that many are now struggling to manage. This sudden inundation has been called “code overload.” Reportshave shownthat AI-generated code can introduce significant problems—including bugs and other quality issues—into codebases, which then have to be fixed by senior engineers before the code can be “shipped” to market. Now, a new company is trying to fix this problem using the same tool that caused it: AI. Gitar—a startup founded by Ali-Reza Adl-Tabatabai, a veteran of Intel Labs, Google, and Uber—emerged from stealth Wednesday with a $9 million funding round led by Venrock, with participation from Sierra Ventures. Two-year-old Gitar sells subscription access to its platform, which deploys AI agents to perform a wide variety of code-quality operations, including code reviews and the management of continuous integration workflows—the automated process of regularly merging and testing code changes to keep a codebase stable and up to date. The platform also gives engineering teams the ability to create their own agents, which can then conduct security and maintenance operations on their behalf. AI-generated code means “more code to review, more tests to write, more CI failures to diagnose,” Adl-Tabatabai, who is CEO, told TechCrunch. What Gitar does is “code validation,” he said—a way to ensure that what’s being built within an enterprise is ready for primetime. “Generation produces code; validation makes it trustworthy. Gitar is the workflow agent that owns that process, orchestrating reviews, tests, and diagnostics end to end,” he added. In the future, Adl-Tabatabai sees automation taking on an even more comprehensive role in software development. “Right now, code that gets shipped into production involves human review, and there are good reasons for that, right? You want to make sure that there’s oversight, and humans are checking to make sure that nothing bad is being shipped.” His vision is that human code reviews will become a minimal part of the process, with companies instead trusting Gitar’s platform to handle those tasks and ship faster. “We have a validation agent that can automatically ensure that your code is safe to ship, and involves humans only in exception cases,” Adl-Tabatabai claimed. There are already numerous other companies already operating in the automated code-review space, but Gitar hopes to distinguish itself through its singular focus on the problem. “Most of the market chased [code] generation. We didn’t,” Adl-Tabatabai said. “Gitar is built around what happens after code is written.” The new funding will be used to hire across Gitar’s engineering and product teams, while the San Mateo company doubles down on developing the systems that allow it to provide its services at scale.
View
Adobe’s new Firefly AI assistant can use Creative Cloud apps to complete tasks
Last October, Adobe previewed a new assistant underthe “Project Moonlight” monikerthat could do tasks for you by tapping different Adobe apps like Acrobat, Photoshop, and Express. That product is now being launched as Firefly AI Assistant. The Firefly AI Assistant will be available in public beta in the coming weeks. The company didn’t specify if the AI assistant will be priced differently from Firefly’s credit-based subscription tiers. Like other creative tools, the Firefly AI Assistant lets you describe what you want it to create, and it will handle the rest. Adobe says the assistant can work across apps like Firefly, Photoshop, Premiere, Lightroom, Express, Illustrator, and its other apps to do tasks for you. Users can control the outputs from the AI assistant using text prompts as well as buttons and sliders. The assistant can suggest actions, orchestrate between actions and apps, and execute workflows, but leaves room for users to interject at any time. The agent will also show controls based on the project you are working on. For instance, if you’re editing a product photo set in a forest, the assistant might give you a simple slider to increase or reduce the amount of trees and foliage. Adobe says the assistant will learn more about your creative preferences over time and suggest actions accordingly. Adobe is also releasing skills, which consist of multiple steps, for the assistant. The “social media assets” skill, for instance, can help you adapt images to different platforms by cropping or expanding, optimizing file sizes, and storing the outputs. Loading the player… Adobe has been working steadily to launch AI-powered assistants forPhotoshop, Express, andAcrobat. The company said on Wednesday that it is exploring having these assistants work better with third-party large language models. Other competitors likeCanvaandFigmaare also working on agentic workflows, though Adobe says its strength is in unifying its existing and popular tools. “We have the opportunity with the Firefly AI assistant and with agentic experiences to remove some of the friction in learning this large catalog of tools we have and bring all of that value to our customers at their fingertips. And that’s the opportunity we have,” Alexandru Costin, vice president of AI and innovation, creativity and productivity business at Adobe, told TechCrunch. Adobe is adding new features to the Firefly tool, too. The AI video editor is getting an option to reduce noise in speech, adjust reverb and music, a color adjustment tool, and now integrates with Adobe’s stock library. The company is also adding the Kling 3.0 and Kling 3.0 Omni models to Firefly’s library of third-party AI models.
View
Reid Hoffman weighs in on the ‘tokenmaxxing’ debate
Days after Metashut downits internal “tokenmaxxing” dashboard following news of the AI leaderboardleaking to the press, LinkedIn co-founder and venture capitalist Reid Hoffman came out in support of the concept that’s recentlytaken Silicon Valley by storm. An AI token is a small chunk of data that an AI model processes when it’s trying to understand a prompt and generate a response. It’s also the unit that’s used to measure AI usage and determine how much AI services cost. As a result, many companies have begun internally tracking which employees are using the most tokens as a proxy for understanding those who are more readily embracing AI tools. They’re calling this concept “tokenmaxxing” — the “maxxing” being Gen Z lingo for optimizing something, as you may have heard in other slang, like “looksmaxxing” or “sleepmaxxing.” However, engineers at tech companies have been arguing whether or not this metric is a viable measure of productivity in the workplace, as it’s akin to ranking people based onwho spends more moneythan others. .@johncoogansays the recent reporting on Meta's 'tokenmaxxing' is less of a sign of bad incentives at the company, and more of a tell about its potential strategy for more vertical integration:"I think it makes clearer the strategy with MSL. Because it's clear that they're…https://t.co/osZD8c6JT3pic.twitter.com/mjh46Diwei Leaderboards that celebrate employees by how much they use AI are sparking debate—critics call it the wrong metric, while supporters say “tokenmaxxing” is critical for mastering the AI agehttps://t.co/ZBHZSWrQ3L Hoffman, in an interview aired at Semafor’s World Economy summit this week, offered his advice for companies adopting AI, saying he had a favorable view of the practice. Though he didn’t refer to the metric in Gen Z-speak, he did express that tracking employee token spend was a good idea. “You should be getting people at all different kinds of functions actually engaging and experimenting [with AI],” Hoffman said at the event. “Here’s one of the things that is a good dashboard to be looking at — doesn’t mean it’s a perfect example of productivity, but… how much token usage are people actually doing as they’re doing it?” He went on to explain that some people may be using a lot of tokens, but in more random or exploratory ways, which is why you want to pair tracking the “tokenmaxxing” practice with an understanding of the things people are using their tokens to do. “Some of it will be experiments that’ll fail — that’s fine. But it’s in that loop, and you want a wide variety of people using it essentially, collectively, and simultaneously,” Hoffman added. Hoffman shared other advice to companies trying to figure out their AI strategies, too, suggesting that AI should be embedded across the entire organization. He also suggested regular check-ins to share what works with others. “We should have, essentially, a weekly check-in. It doesn’t have to be everyone, all the time with each other –but a group check-in about ‘what did we try to do new this week, to use AI for both personal and group and company productivity, and what did we learn?’ Because what you’ll find, some of the things are really amazing,” Hoffman said.
View
After sale of its shoe business, Allbirds pivots to AI
After selling its shoe brand and assetslast month for $39 million, Allbirds is pivoting to AI. Of course, the company is also changing its name, since the footwear brand “Allbirds” was part of the sale. Introducing: NewBird AI, a “fully integrated GPU-as-a-Service and AI-native cloud solutions provider,” the companyannouncedvia its investor relationssiteon Wednesday. The rebranded AI company also announced a $50 million investment from an undisclosed institutional investor in the form of a convertible financing facility. It’s objectively pretty funny that Allbirds is becoming an AI company — not because it’s unusual for companies to pivot, but because of how extreme this pivot is. The maker of the shoesonce craved by the Silicon Valley tech setis now going to be a provider of GPUs. It’s somewhat absurd — and risky — but you can see how the business came to this decision. After the asset and brand sale, Allbirds can keep the public company’s shell (it’s been traded on Nasdaq under the ticker symbol “BIRD”) and then reuse it to invest in the hot AI sector. This recalls the time in 2017 when theLong Island Iced Tea company pivoted to the blockchain, prompting the stock to jump some 275% after the rebranding. That pivot didn’t pan out, as theNasdaq stock exchange delisted the stockthe following year after bitcoin fever died down. Allbirds-turned-NewBird is likely hoping for a different outcome. The company says that the financing and the asset sale are still subject to stockholder approval, with a meeting planned to take place on May 18. If the sale goes through, stockholders will receive a dividend during the third quarter. The new owner of the Allbirds brand and assets, American Exchange Group, will continue to make products for Allbirds customers. Meanwhile, NewBird AI plans to use the new financing to acquire GPU assets, which it will offer to customers seeking AI compute capacity. Over time, the company hopes to grow its service offerings through partnerships and even strategic mergers and acquisitions — if the opportunity arises.
View
Indian IT Majors Report Rise in Sexual Harassment Complaints Since FY21
TCS has attributed higher reported cases to increased employee awareness and reporting.
View
Karnataka Prepares an Ambitious Quantum Roadmap to Secure 20% of Global Market
The state has raised its quantum fund by 17x and named two new research institutions.
View
Karnataka to Unveil Global Innovation Alliance 2.0 for Deeper Tech Partnerships
Representatives from 80+ countries will attend the April 17 conclave in Delhi as Karnataka pitches itself as a gateway for global innovation into India.
View
Axtria Acquires Conexus Solutions to Build AI-Powered CRM Stack for Life Sciences
Axtria said the combined offering will enable life sciences companies to move from fragmented workflows to an integrated model spanning data, analytics, CRM, and field execution.
View
Meta Partners With Broadcom to Develop Next Generation of Its AI Chipsets
Meta announced a new partnership with semiconductor giant Broadcom on Tuesday. The collaboration is focused on developing multiple generations of custom silicon for Meta's artificial intelligence (AI) apps and services. Currently, the company powers its AI compute via the Meta Training and Inference Accelerator (MTIA) chips. The Menlo Park-based tech giant recently shared that in the next two years, it plans to release four new generations of its MTIA chips. It is likely that all of them will be developed in partnership with Broadcom.
View
Google’s SynthID AI Watermarking Tech Claimed to Be Reverse-Engineered
In the world of artificial intelligence (AI) detection, Google's SynthID stands out. First previewed in 2023, it is a watermarking technology that adds an invisible and imperceptible layer of watermark into the content, including text, images, videos, and audio. The Mountain View-based tech giant claims that SynthID cannot be removed, but a developer is now claiming that he was not only able to reverse engineer the watermarking tech, but also distort it to the level that it confuses systems detecting AI content.
View
Enviar tu herramienta
PoweredByAI.app es un directorio de herramientas de IA que ayuda a personas, empresas y creadores a descubrir las mejores herramientas de IA para escritura, programación, diseño, productividad y más.
© 2026 , Producto de011BQ. Todos los derechos reservados.