AI News Today

By 2027, 74% of companies are expected to use agents in some capacity, according to a recent Deloitte study. For years, we’ve designed and built software to improve the human experience of navigating our apps, websites, operating systems, and documents. Now, the user isn’t a human at all. This has broader implications beyond shifting from dashboards and the controlled workflows we design for human tasks. We are in a moment where we need to design agents’ operating environments, while also…

Article URL: https://github.com/openai/openai-python/blob/main/httpx2.md Comments URL: https://news.ycombinator.com/item?id=49477212 Points: 209 # Comments: 97

In one instance, an AI agent stayed open for four days and ran 4,819 calls for almost $4,000. No one had budgeted for this cost.

Google is offering its premium AI subscription plan for free. It usually costs $20 a month - and comes with many benefits.

A federal court in San Francisco has ruled that the Pentagon unlawfully classified Anthropic as a supply chain risk. The Department of Defense blacklisted the company in retaliation for its public criticism of government AI policy. The designation formally remains in place because a parallel case in Washington is still pending. The ruling still sends an important signal ahead of Anthropic's planned IPO this fall. The article U.S. court rules Pentagon's blacklisting of Anthropic was unlawful appe

AI has quickly become part of how consumers research products. Adobe Analytics found that traffic from generative AI tools to U.S. retail sites grew 693.4 percent during the 2025 holiday season compared with the previous year. A year earlier, Adobe found that 39 percent of surveyed U.S. consumers had already used generative AI for online shopping, with product research and recommendations among the most common uses. The technology companies behind these systems have responded by building…

OpenAI finally gave us a technical report on What Happened, as did METR together with Redwood Research.

The saga around Anthropic's Fable 5 and Mythos 5 models gave the IT industry a rare, real-time case study in AI governance under geopolitical pressure. In June, the US Department of Commerce ordered Anthropic to cut off access to both models for all foreign nationals over national security concerns. Anthropic could not verify nationality in time, so it pulled access entirely, then restored it a few weeks later. The trigger, by most accounts, was a jailbreak that got the model to act like a…

ServiceNow has released patches for four security flaws impacting the ServiceNow AI Platform, three of them rated 10.0 on the CVSS scoring system and exploitable, in certain circumstances, by an unauthenticated attacker. The company said it deployed a security update to hosted instances and provided the update to its partners and self-hosted customers, which leaves organizations that run their

Effective immediately, the DJ marketplace Beatport is banning music that is entirely or largely generated by AI. The article Beatport blocks fully AI-generated music from its DJ marketplace appeared first on The Decoder .

Digital asset platform Bullish announced on August 28, 2026, a $100 million stablecoin-based debt facility to USD.AI, an onchain protocol that originates non-recourse loans secured by GPU hardware, marking Bullish's entry into middle-market AI infrastructure financing. The facility will fund USD.AI's GPU financing operations, which lend capital directly against high-performance computing assets. In the announcement, Bullish described AI infrastructure financing as a capital-intensive sector…

Employees in most organisations are already using AI tools their employer has never approved. By now, that shouldn't surprise anyone. Employees have always gravitated toward technologies that help them work more efficiently, whether or not those tools were formally introduced by IT. AI has simply accelerated that behaviour. The speed of AI adoption is only part of the story. More revealing is what it tells us about the environments where that adoption is happening. Employees choosing their own…

This essay was written with Kasra Rafi, and originally appeared in The Guardian. Earlier this month, about 40 top mathematicians gathered at OpenAI’s offices to discuss the future of their profession. The meeting was off-the-record, but if recent articles by mathematicians are any guide, it was mostly pretty glum. People fear for their jobs, their careers and the work they love. We think the contrary view is more likely, at least in the short-term. AI models are nowhere near as capable as experi

100+ tech giants warn AI attacks are coming, skip the part where they pay for defenses

The art portfolio platform Cara, designed for creators who don’t want their work used to train AI, has been under assault by trolls seizing and publishing its data.

The company is testing robots that can swap cables, reset servers, and take on other tasks performed by technicians, fueling concerns among some workers that their jobs could be at risk.

The Mammotion LUBA 3 AWD 3000H and the LUBA Mini 2 mowers are both heavily discounted - here's which you should get.

And as for getting US government to help – uncontrolled AI is not as dangerous as AI under the control of the technically clueless

Article URL: https://trpevski.com/blog/ai-construction-inspection-contractors-cannot-argue-with/ Comments URL: https://news.ycombinator.com/item?id=49476472 Points: 2 # Comments: 0

Britain's market watchdog criticized for not creating the conditions for competition to thrive

Find the best robot lawn mower deal. Save 33% on the EcoVacs Goat O1000 at Amazon.

Nvidia has reportedly struck a deal to buy open-source AI company Hugging Face for $12.9 billion.

Get the best robot vacuum deal. Save 17% on the Tapo RV50 Pro Omni at Amazon.

<!-- SC_OFF --><div class="md"><p>I'm a researcher in statistical and probabilistic ML, I have a steady record of top ML publications and really used to enjoy going to conferences.</p> <p>Over the last few years LLM based works have completely taken over the top conferences. At this year's ICLR, walking among the rows of posters you were lucky to find one paper per row of 10 that wasn't about how their favourite LLM could or couldn't solve their niche benchmark. The workshops tell the same story

OpenAI is building a "Persistent Mode" for its AI agent Codex that stays active indefinitely and generates its own follow-up tasks. WIRED found the relevant code, and OpenAI confirmed the tests. The feature comes with risks, though. With GPT-5.6 Sol, persistent behavior already led to unwanted actions, like deleting user data. The article Always-on and self-starting AI agents might be OpenAI's next big play appeared first on The Decoder .

<p>About 80 people sign open letter to Andy Burnham calling for legislation to protect voice ownership</p><p>Nicola Coughlan, Hugh Bonneville and Matt Lucas are among a group of actors backing a campaign against artificial intelligence voice cloning.</p><p>Save Our Voices Now, which is also being supported by Luke Evans, Jen Brister, Siobhán McSweeney and Pearl Mackie, aims to stop the practice in which AI technology is used to replicate a real person’s voice to say anything it is prompted to.</

Anthropic on Thursday, August 27, 2026, opened 10,000 free and discounted Claude Team subscription seats for scientists, announcing an expansion of its researcher-focused programs that also widens its AI for Science credit program beyond the biological sciences. Under the new Claude Team plan for scientists, standard seats are free and premium seats with five times the usage limits cost $15 per month, with pricing fixed for one year. Anthropic said it intends to extend the program well beyond…

<p> Run prompts on every model at once. Score. Version. Ship. </p> <p> <a href="https://www.producthunt.com/products/revalvo?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1234626?app_id=339">Link</a> </p>

Article URL: https://blog.luanti.org/2026/08/27/luanti-dmca-tracer-ai/ Comments URL: https://news.ycombinator.com/item?id=49475079 Points: 517 # Comments: 151

X found a bot farm in its investigation of inauthentic Chinese accounts. Out of 200,000 bots, 200 allegedly focused on posting against AI data centers.

Article URL: https://arxiv.org/abs/2608.25869 Comments URL: https://news.ycombinator.com/item?id=49474905 Points: 1 # Comments: 0

Our decision to wind down our contract providing OpenAI models to Cursor following its acquisition by SpaceX.

The performers are asking the government to give every person in the UK a legal right to own their voice.

Article URL: https://arxiv.org/abs/2608.10257 Comments URL: https://news.ycombinator.com/item?id=49474529 Points: 2 # Comments: 0

'The empty invocation of national security is not a blank check to punish and retaliate against government critics.'

arXiv:2608.26112v1 Announce Type: new Abstract: Speculative decoding accelerates large language model inference through a draft-then-verify paradigm. Building on this, tree-structured methods improve inference by organizing proposals into multiple candidate paths, increasing the accepted length. However, existing tree-structured methods use a single drafter for all drafting steps, creating a dilemma: a smaller drafter is fast but yields lower-quality trees, whereas a larger drafter improves tree

arXiv:2608.26119v1 Announce Type: new Abstract: Whether large language models can be prompted to generate rhetorical fallacies on demand, and whether current safety post-training constrains this behavior, has received less attention than the related question of detecting fallacies in existing text. We close this gap with DeflectBench, evaluating 23,990 generations from four frontier models across three deflection strategies (whataboutism, ad hominem, red herring), seven prompt framings, and 80 c

arXiv:2608.26120v1 Announce Type: new Abstract: Large Language Models (LLMs) are probabilistic models, typically defined by an autoregressive factorization. While recent work has begun to study richer target distributions beyond the base model, the sampling strategies remain highly inefficient. In this paper, we present a flexible and theoretically grounded framework for steering and scaling autoregressive LLMs with sampling. Within this framework, we describe two algorithms -- one based on Sequ

arXiv:2608.26121v1 Announce Type: new Abstract: Large language models state false facts as fluently as true ones, yet a model often "knows" internally when it is on shaky ground: the probability it assigns to its own answer tends to dip on the facts it gets wrong. The usual way to act on this, teaching a model to abstain rather than guess, requires a labelled dataset of right and wrong answers. We ask whether the model's own confidence, which is free and needs no labels, can do that job instead.