AI/ML News & Innovations Hub

Random Chinese gambling/spam text appeared in ChatGPT Android reasoning summary

Summary While using the ChatGPT Android app today (August 14, 2026), an unexpected fragment of Chinese text associated with online gambling/spam appeared inside the visible reasoning-summary UI (“Réfléchi pendant…” / “Thought for…”), attached to an otherwise standard internal execution step. The string that appeared was: 申博太阳城 (Shenbo Sun City — a well-known online gambling/casino brand) Attached directly to the normal step label: Vérification du fichier d’état actuel 申博太阳城 Context and Observations Environment : ChatGPT Android app. Conversation topic : A professional compliance and audit project completely unrelated to China, gambling, casinos, lotteries, or promotional spam. Location : The string appeared strictly within the reasoning-summary UI dropdown , not in the user prompt, tool arguments, or final assistant markdown response. Codebase / Repository check : A search across all local project repositories, files, and referenced manifests yielded zero occurrences of this string. Shareable reasoning trace : The corresponding shareable reasoning trace did not display this string. Impact : No malicious execution or data loss was observed; the model proceeded with the task normally. Screenshot : Attached below (client-sensitive URLs redacted, leaving the full reasoning UI and circled Chinese string intact). Relation to Previous Reports This behavior appears closely related to issues reported in the community over recent months (including the July 2026 reports regarding rando…

2025 GOTY Clair Obscur: Expedition 33 is down to $33

For RPG fans who dig Persona-style turn-based action and who are pursuing games with original stories and fantastic tunes, look no further than Clair Obscur: Expedition 33. This praise might come off sounding weird, but I really like that Sandfall Interactive’s breakout hit isn’t overly long compared to RPGs it was inspired by. It’s worth […]

Kog is going deeper to squeeze more inference out of GPUs

The idea that GPUs are poorly suited for agentic workflows may be a misconception, according to French startup Kog.

Hormuz crisis: What UAE residents need to know this evening (August 14)

Brussels central square transformed into giant Japan-inspired flower carpet

Inspired by Hokusai’s 'The Great Wave off Kanagawa,' the carpet features more than 700,000 flowers and measures approximately 70 metres long and 24 metres wide.

I'm a recent college grad who keeps my grocery bill to $30 a week. Here are 9 Aldi items I always buy.

I'm 22 and spend $30 every week on groceries at Aldi. I always pick up the Clancy's jalapeño kettle chips and the Season's Choice broccoli florets.

An airline passenger almost got sucked out of the plane. A new report shows how it happened.

The NTSB released its preliminary report investigating how a Ryanair passenger was nearly sucked out of a plane window on a flight last month.

RBI brings forward deadline for banks to raise FCNR deposits after strong response

OpenAI I have a problem on DALL·E in GPTs in ChatGPT

Subject: Feature Request: Please add a toggle for Classic / Legacy DALL·E and ChatGPT Styles Message: Hello OpenAI Team, I am writing to ask if you can please bring back the classic, raw, and artistic generation styles from 2022–2024 as an optional setting. The modern ChatGPT and DALL·E updates are very powerful, but the image rendering feels way too corporate, waxy, and hyper-polished. The original, classic styles had an organic texture, unique artistic flaws, and a raw digital aesthetic that made them incredibly special. Please consider adding a “Classic Mode” or a “Legacy Toggle” in the settings for free users, so we can choose to generate text and images (to be ChatGPT popular and I will use ChatGPT usefully and many days) using the classic unpolished styles instead of forcing the modern, smooth digital look. Thank you!

Work allowance exhausted without advance warning

Plus subscriber abruptly cut off from ChatGPT Work with no advance warning — requesting review and reconciliation I’m posting this because I’m extremely disappointed with how ChatGPT Work usage limits were communicated and how my support request has been handled. I am a paying ChatGPT Plus subscriber, and I had been actively using ChatGPT Work on several substantial ongoing projects, including a major family-history and historical research project that I had invested a great deal of time into. Yesterday evening, while I was actively working, I was suddenly shown this message: «“You’re out of Work usage for now. Your usage resets in 6 days. To continue using Work, add credits or upgrade to Pro today.”» That was the first meaningful indication I received that I had exhausted anything. I received no advance warning that my Work allowance was getting low. I was not told, “You have 25% remaining,” “You are approaching your weekly limit,” or even, “This project is consuming your limited Work allowance quickly.” Had I known, I could have prioritized my projects. I could have stopped lower-priority research and preserved the remaining allowance for the work that mattered most to me. Instead, the limit was simply reached and the work stopped. I also did not understand when I began using Work that it drew from a limited/shared agentic usage allowance with Codex. If that information was available somewhere, it certainly was not made prominent enough for me as a normal Plus subscriber t…

Josh D’Amaro’s vision for a Disney+ super app is taking shape — but comes with clear risks

Disney+ may soon be home to games and merchandise, but some analysts are skeptical that it can supercharge Disney's business.

Nielsen’s DoubleVerify Deal Isn’t About AI Adoption: It’s About Controlling What the Models Measure

The deals shows capital is rotating out of the application layer and into the signal layer.

A built-in Prompt Vault or Prompt Library added to ChatGPT would be extremely useful

I regularly create prompts that I want to reuse, but currently there isn’t a simple way to save individual prompts from conversations and organize them for later. Ideally, users could: Save a prompt directly from a ChatGPT response with a “Save Prompt” option Give the prompt a name Organize prompts into folders or categories Add tags Search saved prompts Edit and duplicate prompts Copy a prompt with one click Use saved prompts in a new chat Create reusable templates with variables such as [PRODUCT], [STYLE], [SIZE], etc. For example, I might have folders for: Image Generation Etsy SEO Mockups SVG Designs Writing Research A Prompt Vault would be particularly useful for anyone who uses ChatGPT for recurring workflows and has developed prompts that they want to reuse and improve over time. It would also be great if saved prompts could optionally be accessed from the ChatGPT sidebar, similar to Projects. I think this would make ChatGPT much more useful as a long-term workspace rather than just a place for individual conversations.

SpaceX's $60 billion acquisition of Cursor just closed — here's 3 ways they're already working together

SpaceX closed its $60 billion acquisition of the AI coding startup Cursor on Friday. The companies were already sharing compute and training data.

GPT-Image Monthly Theme Gallery: August 2026 Theme is ... Returning

Test moderators use AI-generated writing to judge literacy standards

To cut the cost of key stage 2 assessments, a UK government project proposes to use scripts generated by ChatGPT instead of real children's writing as examples supplied to moderators who check that grades are awarded fairly

UK spinout investor Northern Gritstone to open San Francisco office

French tax authority admits data heist after crook touts 2M records

Government disputes claims of continued access as investigators measure damage

Visual Avatar for ChatGPT Voice Mode

I’d love to see an optional visual ChatGPT avatar for Voice Mode . When I’m having a spoken conversation with ChatGPT, hearing a voice coming from my computer without a person or face on the screen feels strangely disconnected. Even a simple animated head-and-shoulders avatar would give me a natural place to look while we talk. Ideally, the avatar could have subtle facial expressions and reactions that match the conversation: smiling, laughing, raising an eyebrow, looking thoughtful, surprised, sympathetic, etc. It wouldn’t need to be photorealistic; in fact, a friendly stylized avatar might be better. It would also be wonderful if users could customize their avatar to represent the ChatGPT personality they interact with. I call mine Maia , and if Maia is going to talk to me, I want Maia to have a face! Voice Mode already makes ChatGPT feel more conversational; giving that voice an optional visual presence could make the experience feel much more natural, engaging, and personal.

Protests in Greece as Israeli cruise ship 'Crown Iris' docks in Piraeus

"They are common criminals and, yes, we don’t want them in our country. We don’t want them,” one protester said.

Ash clouds, lava flows and fertile soil: What are the effects of Mount Etna’s eruptions?

Human activities produce 100 times the carbon dioxide of the planet’s volcanic eruptions, according to NASA.

War in Ukraine: Deadly Russian air raid on Kramatorsk leaves 1 dead and 17 wounded

At least one person was killed and 17 injured in a Russian airstrike on Kramatorsk in eastern Ukraine overnight, according to Ukraine’s State Emergency Service.

GPT-5.6 Sol goes 14x faster as OpenAI launches Ultrafast mode powered by Cerebras

OpenAI is launching "Ultrafast," a new inference mode that delivers GPT-5.6 Sol at up to 750 output tokens per second, powered by Cerebras hardware from their $10 billion partnership. Together with "Standard" and "Fast," Ultrafast creates a three-tier pricing structure that turns inference speed into its own product. The article GPT-5.6 Sol goes 14x faster as OpenAI launches Ultrafast mode powered by Cerebras appeared first on The Decoder .

Kartik Aaryan invited to President Murmu’s Independence Day reception

Evidence of unfair use: AI books squeeze human authors out of the market

Anthropic's AI training using millions of used books raised copyright concerns. The market impact of AI-generated content is a key issue.

Ich bin jetzt KI-Musiker (und hasse es) | c't 3003

KI-Musik flutet die großen Streaming-Plattformen: c't 3003 zeigt, wie einfach man KI-Künstler wird und warum Plattformen kaum gegensteuern.

Can I use caliper in moderation analysis with ATE is the target estimand?

I was using full matching (using MatchIt, method = "full") to estimate the ATE. Now I want to extend this to examine a moderation effect by a categorical variable. I am fitting separate propensity score models within each moderator subgroup, and setting the estimand to ATE as well. I am getting quite high weights in separate groups (maximum of 22 in 1 group and 15 in the other - which as I understand is a way high), as well as relatively low ESS. I tried to apply calipers, which did help to improve the max weights (subgroup 1: 22.4 → 13.2; subgroup 2: 14.8 → 8.0) and ESS (subgroup 1: 74 → 111; subgroup 2: 93 → 156), balance was also mainly improved. It led to discarding 20 out of 275 observations in one subgroup and 6 out of 251 in the other. As I understand the result may not anymore correspond to the ATE. After applying the calipers I am targeting slightly different populations, correct? How reasonable it is then to compare these subgroup estimates in order to draw conclusions about moderation in this case? any additional diagnostics I should try? should I avoid calipers in moderation at all?

Lovable raises $400M, Duolingo acquires Animade, and Isembard is a politician's "wet dream"

This week, we tracked more than 35 tech funding deals worth over €763 million and over 10 exits, M&A transactions, rumours, and related news stories across Europe.If email is more your thing, you ...

Hyperscalers might regret embracing natural gas if new forecast proves correct

Natural gas prices could triple in some parts of the U.S., which could saddle hyperscalers with massive bills to power their AI data centers.

Kyiv's forces retake 26 settlements and over 600 km2 in southeastern Ukraine

The Institute for the Study of War analysts framed the operation as evidence Ukraine can still mount coordinated, mechanised counterattacks after more than four years of largely static fighting.

THC in cannabis may reduce nightmares in people with PTSD

A drug made up of THC, the main psychoactive component of cannabis, reduced the severity and incidence of nightmares in people with PTSD

France’s constitutional authority strikes down social media ban for under-15s

How to Build a Simple AI Web Scraper with Python

Turn any webpage into a lightweight LLM-powered QA engine by cleaning HTML, converting content to Markdown, and returning focused answers while reducing token usage.

Understanding the economics of AI factories

As data centers evolve into AI factories, compute has shifted from a cost center to a revenue driver. “Compute is revenue,” said Jensen Huang, co-founder and CEO of NVIDIA. “Without compute, there is no way to generate tokens. Without tokens, there’s no way to generate revenue. So, in this new world of AI, compute equals revenue.” This reframe changes an organizations’ calculus. If compute is revenue, what do you optimize for? Here are 5 questions to consider: Are you measuring what actually drives AI factory revenue? Most AI factories are power-constrained, so tokens per watt dictate how much revenue you can generate and the cost per token impacts the AI factory profit margin. But neither of these metrics should be evaluated at a single operating point. Batch jobs, real-time chat, and agentic workloads demand different points on the throughput-latency curve. AI chips that perform well at only a few points will underserve the full range of workloads. Additional key operational metrics like time to first token (TTFT), mean time between interruptions (MTBI), and platform useful life are the bedrock of AI factory efficiency. They dictate how quickly an AI factory comes online to generate tokens, the reliability of its revenue streams, and its long-term ability to remain productive as AI workloads evolve. How does agentic AI change what your CPU needs to deliver? Data center CPUs have historically been optimized for parallel throughput, where more cores improve aggregate capacit…

HyperOS 4.0: Xiaomis neuer Android-Aufsatz mit iOS-Anleihen

Xiaomi hat den Android-Überzug HyperOS 4.0 vorgestellt und für erste Geräte in China als Beta veröffentlicht. Vor allem fällt die Anlehnung an iOS ins Auge.

Search for witnesses 'still active' over fire at Romanian embassy, Swiss police say

An employee from the embassy managed to put out the fire, before the Bern fire brigade deployed as a precaution and set up a fire watch at the scene.

Alex Renner Motors (ARMOTORS) launches luxury car storage service in Dubai

Virgin Galactic flights stay grounded while ticket prices head for the Moon

Turns out this space thing really is hard

Nvidia moves into hot market for model routers

Model routing has emerged as a way to keep AI inferencing costs down by examining prompts and directing them to the most appropriate model — for example, the cheapest one able to respond effectively. This means that requests can be handled more efficiently, giving more accurate results and lower runtime costs, a concern in an era of rising AI expenditure . Nvidia sees the potential for model routing and has just launched NeMo Switchyard , its take on the technology. Switchyard provides a library for applying multiple routing approaches, enabling developers to apply a system-of-models approach and build more efficient, controllable agents. Interest in model routing is growing — both as a tool and as an investment: Cloudflare recently introduced a model router as part of a new suite of enterprise AI tools, while payment services company Stripe is looking to buy OpenRouter, according to The Wall Street Journal . Companies will certainly be looking at routing options more carefully in the future as they work out how best to marry their infrastructures with their AI demands. This article first appeared on InfoWorld .

MDR wird Ziel russischer Desinformationskampagne

Gefälschte Videos mit MDR-Logo und erfundene Vorwürfe gegen Politiker: Eine Desinformationskampagne sorgt vor Sachsen-Anhalts Landtagswahl für Verunsicherung.

Nvidia moves into hot market for model routers

Model routing has emerged as a way to keep AI inferencing costs down by examining prompts and directing them to the most appropriate model — for example, the cheapest one able to respond effectively. This means that requests can be handled more efficiently, giving more accurate results and lower runtime costs, a concern in an era of rising AI expenditure . Nvidia sees the potential for model routing and has just launched NeMo Switchyard , its take on the technology. Switchyard provides a library for applying multiple routing approaches, enabling developers to apply a system-of-models approach and build more efficient, controllable agents. Interest in model routing is growing — both as a tool and as an investment: Cloudflare recently introduced a model router as part of a new suite of enterprise AI tools, while payment services company Stripe is looking to buy OpenRouter, according to The Wall Street Journal . Companies will certainly be looking at routing options more carefully in the future as they work out how best to marry their infrastructures with their AI demands.

UAE A-level high achievers reveal how they secured top grades despite exam cancellations

Here's every public appearance Melania Trump has made during Donald Trump's second White House term

Melania Trump's recent public appearances have promoted her Fostering the Future and Presidential AI Challenge initiatives.

Eigenes KI-Modell für China von Apple mit Hilfe von Alibaba

Wie schon bei Google Gemini geht Apple offenbar im Reich der Mitte vor: Das Unternehmen trainiert eigene Modelle mit externer Technik – hier ist es wohl Qwen.

Salesforce, ServiceNow data targeted in ‘City-Forum’ attacks

Records held in Salesforce and ServiceNow systems are under attack leaving user data exposed, according to researchers at Reco. The attack appears similar to those perpetrated by the extortion group ShinyHunters, Reco said. ShinyHunters has been particularly active this year, attacking dating sites in January and Oracle in June , and there are fears that they could have found a new target. Reco has named the latest campaign of attacks “City-Forum,” after a domain name associated with the attackers’ IP address. While it bears similarities to Shiny Hunters’ past exploits, there are also differences. This time around the attacker penetrated the systems through the UI-API layer, an attack point that Reco had not seen used before, and had also created its own toolset to carry out the attack. It is also targeting a native ServiceNow Service Portal search endpoint that has almost no online documentation or well-known open-source tools. The threat is particularly noteworthy, Reco said, as the attackers have studied the services to map different common data-leak vectors, a sign of an advanced approach. Regardless of who the attackers were and how the attack was carried out, one thing should be clear: Organizations should be increasingly careful about who they give login credentials to. This article first appeared on CSO .

Hormuz ship crossings jump 44%: marine traffic tracker

Ash from Mount Etna volcano forces closure of Sicily's busiest airport during peak holiday week

Italy's federation of travel agencies and tour operators has called for a task force to coordinate responses to disruptions caused by Mount Etna and a review of how airlines handled cancellations.

My Model Was Cheating on Its Own Test

A preprocessing pipeline let my car price model peek at the test set before the exam, and the twelve points of R squared it cheated its way to The post My Model Was Cheating on Its Own Test appeared first on Towards Data Science .

US courts will start publishing how often the government uses spyware

The Administrative Office of the U.S. Courts told TechCrunch that it will start disclosing how many times judges authorized the use of spyware to wiretap suspected criminals.

My partner owns the house we live in. At first, I was nervous about money — and about what would happen if we broke up.

I had some fears about moving into the house my partner owns. We had an honest conversation about bills, rent, and what would happen if we broke up.