GPT-Image - General AI Gallery: A Place for Images with AI created by Users
Some of my latest RPG Images
Latest AI/ML News
35407 matching items
Operational continuity between long-running ChatGPT conversations
When a ChatGPT project lasts weeks or months, the problem is not always that the model simply “forgets.” Something more dangerous can happen: a new conversation reconstructs an almost-correct version of the project and continues reasoning perfectly well from incomplete history. We ran into this ourselves. We were about to change part of a project when we recovered an older continuity record and discovered that: a decision we thought was new had already been made, a capability we thought was missing already existed, and something we were about to modify was supposed to remain frozen. The reasoning wasn’t bad. The available context was incomplete. So we built a simple process called Continuity that extracts the current operational state of a conversation and carries it into a new one: active objectives, decisions still in force, constraints, completed work, unresolved items, dependencies and critical references. It doesn’t try to make the AI remember everything. It preserves what is still necessary to continue working correctly. After repeated testing on long-running projects, the effect has been surprisingly large for such a simple mechanism. This seems like a useful native ChatGPT capability: not unlimited memory, but explicit operational continuity between conversations . There is a free working implementation here for anyone who wants to test the idea: lasserlab.mx Continuity I’d be interested to know whether others working on long-running ChatGPT projects have encountered…
Nigeria’s Mathesis Analytics Raises Strategic Investment From Sewa Capital to Scale AI Credit Decisioning
Nigerian fintech Mathesis Analytics has secured a strategic investment from Sewa Capital to scale its AI-powered credit decisioning engine, bringing a platform that has already supported more than 8 million loans deeper into institutional lending. The investment amount was not disclosed. Mathesis addresses what its founders identify as a structural gap in Nigerian credit markets: [...]
Croatia wildfire near Omiš injures residents and disrupts holiday travel
A major wildfire near Omiš, around 25 kilometres south-east of Split, forced about 1,200 residents and tourists to evacuate overnight on Thursday and into Friday as strong winds drove the flames towards the coastal town.
Kenya Records World’s Highest AI Usage at 97.5% — And Why Another Major Report Puts It at 8.7%
Kenya has the world’s highest rate of AI use, with 97.5% of online adults reporting they used at least one AI tool in the past month, according to the Digital 2026 Mid-Year Global Update published by DataReportal in partnership with We Are Social and Manochi. That is the highest figure among the 54 countries tracked [...]
I live in London, where we're experiencing our 5th heat wave. I have no AC, so here's how I'm surviving my workday.
The UK is going through its fifth heat wave this summer, making work, commuting, and dinner nearly impossible for me. I found some ways to stay cool.
Alibaba Opens Qwen3.8-Max Weights, Letting Teams Self-Host a 2.4T Model
Alibaba drops open weights for Qwen3.8-27B and the massive 2.4T-A95B Max model, both free under Apache 2.0
At least 10 people in intensive care and thousands flee intense wildfire in Croatia
During a visit to the area Prime Minister Andrej Plenković stressed the dangers of the fire and praised the efforts of the firefighters.
US threatens EU over its green reporting rules
Washington says Brussels still hasn’t delivered on its commitments from last year’s transatlantic trade deal.
A Queensland man enjoyed soaring profits from a crypto trading app. Then his money started disappearing
The administrative work required to run an effective investment scam has been drastically cut by using merging technology Get our breaking news email , free app or daily news podcast Within days of clicking on an online ad for a cryptocurrency trading app, a 29-year-old Queensland man watched his dashboard soar with profits. Making investment gains was as simple as downloading the slick-looking app and browser extension linked to his crypto wallet. So he invested more. Continue reading...
How Uber evaluates AI agents at production scale
A background comment about pizza exposed a failure that Uber’s offline evaluations had missed. The incident helped reveal what production AI agent evaluation actually requires: automatic tracing, living datasets, shared ownership, and a direct connection to product decisions. The post How Uber evaluates AI agents at production scale appeared first on Arize AI .
Tech Visionary Says the Big AI Labs Don’t Get What People Want
Tim O’Reilly built a publishing empire that AI is helping to destroy. Yet he loves AI—as long as it’s open source.
RAG Workflow and Loop Engineering: The Dispatcher That Decides When to Loop and When to Stop
Enterprise Document Intelligence [Vol.1 #13] - Putting the patterns together, and why this is what “agentic RAG” should look like The post RAG Workflow and Loop Engineering: The Dispatcher That Decides When to Loop and When to Stop appeared first on Towards Data Science .
MCP OAuth succeeds (token + initialize OK) but ChatGPT web still re-opens auth on tool call
Hi there — we’re closing this topic because there hasn’t been a new response for a long period of time and the discussion appears inactive. If you still need help, please start a new topic with your current question and any relevant, non-sensitive details. Thanks for being part of the community. Avinash
Apple proposes to take a 15% cut of purchases made outside the App Store
Apple is asking a federal judge to allow it to charge commissions of up to 15% on purchases made through external links in iOS apps.
Add Expand/Collapse button to Chat Window in Installed ChatGPT App
Issue For longer messages, the chat window in the ChatGPT app is small making it cumbersome to review what has been typed before hitting send. Feature Request As currently present in the web ChatGPT app, add the ability to expand/collapse the chant window in the installed ChatGPT app. See images below.
I gave up freelancing for a steady paycheck. My first summer as an employed mom is chaos.
After years of freelancing while raising three kids, I took two part-time jobs. I love the steady paycheck, but summer childcare is exhausting.
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.
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.
Comment on The Future of Vision AI: How Apple’s AIMV2 Leverages Images and Text to Lead the Pack by SquareFaceIconGenerator.app
Interesting breakdown of AIMV2. The unified autoregressive prediction over image patches and text tokens is a clever way to get denser supervision, and it makes sense that it outperforms CLIP on multimodal benchmarks. I’m curious how this will translate into real-world tools like image search or even avatar customization — stronger vision encoders could make those noticeably more accurate. Thanks for sharing the details.
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.
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.
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
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 .
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
