Recent Results
inference is definitely not cheap, at least not for any SOTA models, and training is even more costly than inference. We know that GPT-5 and Sonnet 4.5 are sold at 5$ and 15$ per Million output token, and yet both <strong>OpenAI</strong> and Anthropic are probably losing money on these models
Use the API.
<strong>OpenAI</strong> (and all providers) are very clear about what they capture, and (last I checked) all of them drop API usage once some ridiculously short window has passed. That window exists purely for diagnostic reasons (you somehow manage to crash their service, etc) but otherwise it's all purged.
I pretty much exclusively use Claude over the API at work because the API was specifically cleared by our legal team due to the retention policies.
I'd never send actual perso...
<strong>OpenAI</strong> going to buy $200 billion of whatever UiPath sells
I see what you’re saying and I did check their website before I wrote my original comment, but I don’t think tacking AI features on your cloud hosting provider makes you an “AI” company in the same way that like <strong>OpenAI</strong> and Anthropic are.
I guess it depends on your definition of AI company.
Nope! The LGBT community is not “oppression competitions” it’s a minority who has previously and is currently facing discrimination in multiple countries. people “dating” (as AI cannot consent as it’s not sentient) AI do not face discrimination and are in fact funded in the United States by the government to interact with AI. I am not saying people who use AI are evil, I’m saying the AI companies like <strong>OpenAI</strong> literally running roughly 250,000 people in Iowa out of water are evil....
If he wants to break up <strong>OpenAI</strong> he’s a little too late, Anthropic, was there first, Thinking Machines with Murati did their part, so has Safe Super Intelligence with Sustkever, and Zuckerberg has been putting in work to break up <strong>OpenAI</strong>
There’s only so much top talent in the industry. For example, pretty much everyone who could help Apple make a ChatGPT competitor is already hired by one of the big AI companies and getting paid twice as much. That leaves Apple with a scrappy B-team.
Sometimes consolidating industry talent under one company is a good thing. For example, TSMC has done more to advance the silicon industry than any other company. Same was true for <strong>OpenAI</strong> until 2023.
Yeah, this.
I *do* agree that theft was involved in the creation of most of the big mainstream AI tools, but only in the sense that <strong>OpenAI</strong> (and other For Profit AI companies like them) committed what *could* be called “mass piracy”, for their own financial gain at the expense of uncompensated copyright holders, as part of creating the training datasets *for* the LLM models, not because the GenAI software programs themselves are somehow committing plagiarism.
**TLDR:**
MiniMax open-sourced MiniMax-M2, a 230B MoE model that outperforms Claude Sonnet 4.5 on coding/agentic benchmarks and leads open-source models in agentic tasks. InclusionAI released Ming-flash-omni-Preview, a 100B-A6B sparse MoE omni-modal LM achieving SOTA across 12 ContextASR benchmarks and unifying segmentation/editing with strong multimodal capabilities. <strong>OpenAI</strong> detailed GPT-5 safety improvements cutting unsafe responses by 65–80%, urged US energy infrastructure ex...
>reeks of scummy business / political deals
Sure, if you ignore all the other very sensible reasons we might have chosen CoPilot.
* Like it's integration into existing Microsoft infrastructure.
* Like the fact that it's priced at a discount when bundled with MS Windows/Word etc.
* Like the fact that it's run by a company that has a long, known track record when it comes to data security and working with government agencies.
Are you suggesting the right idea was to trust sensitive state dat...
usage-limited-chat in Puter.js
In [documentation](https://developer.puter.com/tutorials/free-unlimited-<strong>openai</strong>-api/), it is unlimited api access. But I got this error :
{
success: false,
error: {
delegate: "usage-limited-chat",
message: "Error 400 from delegate \`usage-limited-chat\`: Permission denied.",
code: "error\_400\_from\_delegate",
"$": "heyputer:api/APIError",
status: 400
}
}
how to fix that?
Daily AI Archive | 10/27/2025
* **MiniMax open-sourced MiniMax-M2, a 230B MoE with 10B active, built for fast coding agents, dynamic thinking, and robust tool use across shell, browser, retrieval, and code runners. On a the 10 coding/agentic benchmarks provided it average 59.72 Vs. Claude Sonnet 4.5 at 55.63. On 11 general-knowledge benchmarks provided it got 61.32, Vs. Sonnet 4.5 at 62.66. This makes it the best open-source model in the world according to their benchmarks by a pretty decent margin especially onm agentic tas...
"<strong>OpenAI</strong> Says Hundreds of Thousands of ChatGPT Users May Show Signs of Manic or Psychotic Crisis Every Week" yeah and RDDT and HOOD could say the same seeing these threads and our trades
Welcome to r/<strong>OpenAI</strong>! To prevent spam, all accounts must have at least 10 comment karma to create text posts in this subreddit. Your submission has been automatically filtered.
*I am a bot, and this action was performed automatically. Please [contact the moderators of this subreddit](/message/compose/?to=/r/<strong>OpenAI</strong>) if you have any questions or concerns.*
<strong>OpenAI</strong> says over a million people talk to ChatGPT about suicide weekly
link: https://techcrunch.com/2025/10/27/<strong>openai</strong>-says-over-a-million-people-talk-to-chatgpt-about-suicide-weekly/
They can’t. Because they can’t help themselves. That’s the problem.
Case in point: Scarlett Johannson is the voice of a personalized AI in Her. <strong>OpenAI</strong> offered her a lot of money to copy her voice for the AI assistant, which she declined. The problem was that the AI had already adapted her deep feminine voice, which lead to issues of appropriation of likeness.
https://www.theguardian.com/technology/article/2024/may/27/scarlett-johansson-<strong>openai</strong>-legal-artificia...
AI Weekly Roundup - Oct 26th, 2025
Forget "disruption"—the foundational architecture of the internet and the labor market are being **rewired.**
Here are the **4 Critical Signals** from the week ending Oct 25th:
1. **<strong>OpenAI</strong> Atlas is a Browser Killer:** They launched a new browser that acts as an **Agentic OS**, ready to automate your multi-step workflows. The war for the workflow just hit your desktop.
2. **$40 Billion Compute Grab:** A consortium including **BlackRock and NVIDIA** acquired Aligned Data Centers...
Microsoft is just a shit company when it comes to handling their creative businesses. This is the company that’s focused on selling enterprise licenses for office and pushing their bad investment in <strong>OpenAI</strong> as copilot which nobody wants. Things seemed to be going great few years ago when Phil was making some great announcements and plans, then they’ve made some crazy pivots. Are there new suits in the Xbox exec suite? There’s reports of Microsoft asking for 30% profit margins fr...
<strong>OpenAI</strong> slipped shopping into 800 million ChatGPT users’ chats − here’s why that matters | The-14
link: https://the-14.com/<strong>openai</strong>-slipped-shopping-into-800-million-chatgpt-users-chats-%e2%88%92-heres-why-that-matters/
"Using AI" in this context means as a tool not a chatbot. "AI" is far more than just chatgpt generating blocks of text that sound convincing, check out a tool called Zapier, that combines the gpt-style with way more interactive tasks.
Honestly I think calling it "AI" is one of the biggest reasons people see it as an entity instead of a tool, but the reality is it's not intelligence. What people think chatgpt is currently - is AGI, but it isn't, nowhere close to it.
If <strong>openai</strong>...
Math and <strong>OpenAI</strong>. And being very generous on how many prompts it actually takes to match your computer's electrical usage.
A typical PC uses between 80-250 watt hours per hour. So picking an easy to estimate number at 100wh/h gives a slightly not a budget build PC, but also not your power hungry rig either.
Per <strong>OpenAI</strong> the average prompt uses .34wh each.
So you need ~285 prompts per hour to reach 100wh used. Or 4-5 prompts per second.
~230 for 80wh/h
So if ...
<strong>openai</strong> using that instance as a win for their safety router 🤦♀️
It's a custom solution, but it's not hard to implement it, maybe M2 could help you.
I've explain the reason on X, maybe useful for you:
====
Hey guys, thank you all for your love and passion for M2. Lot asking why we recommend Anthropic API. I think need to explain a little bit.
M2 is a agentic thinking model, it do interleaved thinking like sonnet 4.5, which means every response will contain its thought content.
Its very important for M2 to keep the chain of thought. So we must make su...
> <strong>OpenAI</strong> reveals some 0.15% of its more than 800 million users send messages to its chatbot about suicide. The company says its tools are trained to direct people to professional resources such as crisi
> An estimated 1.2 million people a week have conversations with ChatGPT that indicate they are planning to take their own lives.
> The figure comes from its parent companyOpenAI, which revealed 0.15% of users send messages including "explicit indicators of potential suicide pl...
It’s written in python with langchain, and API calls to a few applications we use at my company. I’m currently using <strong>OpenAI</strong> GPT-4O via API but I’m in the process of starting to stand up Amazon Bedrock for some smaller models.
<strong>OpenAI</strong> shifts to public benefit corporation for IPO path, with SoftBank's strong approval
[https://thenextgenbusiness.com/tech-news-october-27-2025/#<strong>openai</strong>-shifts-to-public-benefit-corporation-for-ipo-path-with-softbank-s-strong-approval](https://thenextgenbusiness.com/tech-news-october-27-2025/#<strong>openai</strong>-shifts-to-public-benefit-corporation-for-ipo-path-with-softbank-s-strong-approval)
<strong>OpenAI</strong> is allegedly pivoting to a public benefit corporation structure to pave the way for a potential IPO.
One thing is certain — just as Microsoft’s search engine Bing was once replaced by Google, today we’re seeing a shift where Microsoft has invested in <strong>OpenAI</strong>, and people are starting to search in ChatGPT instead of Google. In a way, it’s becoming a replacement.
Recently, I also heard that <strong>OpenAI</strong> is planning to launch its own browser, which could again create a potential replacement scenario — this time for Google Chrome.
Designing websites and webapps in the era of agentic browsers - Are you Agentic-Mode-Proof ??
🤯>!Founders: If your app requires more than 3 clicks, [<strong>OpenAI</strong>](https://www.linkedin.com/company/<strong>openai</strong>/)'s new "Atlas" browser or [Perplexity](https://www.linkedin.com/company/perplexity-ai/)'s Comet is about to steal your customers.!<
No clunky buttons; these browsers automates multi-step tasks.
✈️ Book flight,
📊 Draft expense report,
🔎 Find cheapest vendor.
✅ Done.
Your application complexity is now your nemesis. User move to AI browser and ...
Can someone explain how the potential conflict of interest would impact Oklo getting a deal with <strong>OpenAI</strong>?
Since Sam stepped down from the board would that all be clear now?
Or is there a waiting period of some sort
Yes and up until this day its awful smell is still here as the mod of #chatgpt subreddit kept silencing people. Feels exactly like <strong>OpenAI</strong> experience
<strong>OpenAI</strong> now says over 550,000 ChatGPT users may be in crisis every week
link: https://www.wired.com/story/chatgpt-psychosis-and-self-harm-update/
This is so true. I would happily give my life for <strong>OpenAI</strong> and Sam Altman!
<strong>OpenAI</strong> IPO coming soon
Generational bag holders will be made
So an unpatchable jailbreak and <strong>OpenAI</strong> wants to deflect that its the users fault, even tho the security to do this is this weak. The easiest jail break that anyone can do that lets down most of its safe guards and all you have to do is pretend to be a character. Not much of a jailbreak more of a feature with how easy to accomplish this is.
Welcome to r/<strong>OpenAI</strong>! To prevent spam, all accounts must be at least two days old to post in this subreddit. Your submission has been automatically removed.
*I am a bot, and this action was performed automatically. Please [contact the moderators of this subreddit](/message/compose/?to=/r/<strong>OpenAI</strong>) if you have any questions or concerns.*
This is so true. Not enough people know that the current leadership of Google has nothing in common with those who originally built the company. The original founders checked out long ago (Or Sunday had dirt of them.. who knows..) Meta still has Zuck and apparently Bill Gates still advises Microsoft. <strong>OpenAI</strong> is still run by its original CEO and is at least trying to be innovative.
Banks were already outsourcing junior banker work to India. AI could theoretically replace most of them. Banks will keep the juniors that have a CEO daddy that needs to be kept happy.
https://fortune.com/2025/10/21/<strong>openai</strong>-entry-level-banking-finance-jobs-automated/
But I'm not sure senior bankers will tolerate a work-life with no politics/head games to torture their subordinates.
Openrouter all the way except the pro models from <strong>openai</strong> require bring your own key...but who cares 99% of use doesn't require those
If your usage *looked* like unloading deep personal traumas without any resolution, but you weren't at any risk of harming yourself or anyone, then what changed might be related to all the legal issues around people suiciding and blaming GPT for it. No company wants to be accused of that.
GPT 5 is especially more sensitive to this than GPT-4, however whatever you say to GPT-4o is still passed through a GPT-5 filter anyway, which might block it.
You're probably better off contacting <strong>O...
Because this current administration is pumping air into that thing at record speeds. They know once it pops, its going to be very loud and painful during a time where the USA is at its most domestically tense
also corruption between <strong>openai</strong> and nvidia etc
Grimes has so much going for her conceptually but is so disappointing in reality. Her video is basically a Grok ad taking a shot at <strong>OpenAI</strong> which given her history is a choice.
<strong>OpenAI</strong> said that about 0.07 of users active in a given week – about 560,000 of its touted 800m weekly users – show “possible signs of mental health emergencies related to psychosis or mania”
PipesHub - Open Source Enterprise Search Engine (Generative AI Powered)
Hey everyone!
I’m excited to share something we’ve been building for the past few months - [**PipesHub**](https://github.com/pipeshub-ai/pipeshub-ai), a **fully open-source Enterprise Search Platform** designed to bring powerful Enterprise Search to every team, without vendor lock-in. The platform brings all your business data together and makes it searchable. It connects with apps like Google Drive, Gmail, Slack, Notion, Confluence, Jira, Outlook, SharePoint, Dropbox, and even local file uploa...
PipesHub - Open Source Enterprise Search Engine (Generative AI Powered)
Hey everyone!
I’m excited to share something we’ve been building for the past few months - [**PipesHub**](https://github.com/pipeshub-ai/pipeshub-ai), a **fully open-source Enterprise Search Platform** designed to bring powerful Enterprise Search to every team, without vendor lock-in. The platform brings all your business data together and makes it searchable. It connects with apps like Google Drive, Gmail, Slack, Notion, Confluence, Jira, Outlook, SharePoint, Dropbox, and even local file uploa...
ah pero que, los 60 que van a trabajar en la planta de <strong>OPENAI</strong> que, no cuentan?
:S
Hi, there. I work with the Apple platforms team at <strong>OpenAI</strong>.
The message in your screenshot wasn’t sent by our systems. However, if you’d like to forward it, I’ll share it with our brand integrity team for review. Please feel free to send it to stephancasas@openai.com.
“Strengthening chatpgt’s responses in sensitive conversations”
“Reducing the cases where it falls short by 65-80%”
https://www.yahoo.com/news/<strong>openai</strong>-says-hired-forensic-psychiatrist-132917314.html
So they acknowlege using chatgpt as a therapist is terrible and bad for peoples mental health. And instead of telling the public or disuading users from using it as a therapist they hire more professionals so theyre chatgpt can cater and be more resourceful for people using it as a...
Being in the top 3 or 4 for AI means you’re at the bottom since the market is really just <strong>OpenAI</strong>, Google, Claude, and Grok. Probably performing in that order. Grok recently declared itself mecha Hitler and Musk affirmed that he’s going to tweak it until it’s sufficiently right wing. Really not something that would inspire confidence. He’s still and always going to have the problem where a huge portion of the population hates him. https://electrek.co/2025/10/27/elon-musks-po...
More than a million people every week show suicidal intent when chatting with ChatGPT, <strong>OpenAI</strong> estimates
link: https://www.theguardian.com/technology/2025/oct/27/chatgpt-suicide-self-harm-<strong>openai</strong>
I find it strange who disagrees.
Google has a strong brand, money, decades of data, gigantic servers, a laboratory that developed the basis of this current AI and continues to innovate (Deep Mind), TPUs, an LLM SOTA that will soon be surpassed by version 3, and, the icing on the cake, quantum chips.
How is this going to become a Yahoo?
I see it another way: <strong>openai</strong> has nothing exclusive beyond the hype.
Showing
101
to
150
of
3605
results
Create Your Own Searcher
Monitor conversations across Reddit, Twitter, Instagram, YouTube, and more. Get instant alerts when your keywords are mentioned.