ChatGPT is down

Pivot 5: 5 stories. 5 minutes a day. 5 days a week.

1. Chat GPT is down

ChatGPT, a chatbot developed by OpenAI, experienced an outage on Thursday afternoon when it couldn't respond to queries and some users saw an "internal server error" message. The outage began around 1:30 PM ET and affected ChatGPT, the API, and its text-to-video generator Sora.

Microsoft, OpenAI's exclusive cloud provider, reported a power issue at one of its datacenters, affecting North America and Xbox cloud gaming. Microsoft has since fully restored power to the affected datacenter. ChatGPT has gone down several times in the past several months, including a video generation tool and a widespread outage affecting AI tools in June.

Read the full story here

2. DeepSeek-V3, ultra-large open-source AI, outperforms Llama and Qwen on launch

Chinese AI startup DeepSeek has released its new ultra-large model, DeepSeek-V3, which uses a mixture-of-experts architecture to handle tasks accurately and efficiently. The model, available via Hugging Face, has already outperformed leading open-source models and closely matches the performance of closed models from Anthropic and OpenAI.

DeepSeek hopes these developments will pave the way for AGI, where models can understand or learn any intellectual task that a human can. The new model uses the same basic architecture as its predecessor, focusing on multi-head latent attention and DeepSeekMoE. The company has also introduced two innovations: an auxiliary loss-free load-balancing strategy and MTP, which enhance training efficiency and performance.

Read the full story here

3. You Need to Create a Secret Password With Your Family

Scammers are increasingly using AI to create deepfake videos and voice clones to impersonate people and extract money from victims. The FBI has recommended creating a "secret word or phrase with your family to verify their identity," while British bank Starling has published guidelines on creating safe phrases with others. Fraudsters use manipulation tactics to put victims in a vulnerable state, and having a passphrase or similar prepared strategy can help validate the legitimacy of an unusual interaction and take control.

Companies and families are at risk due to AI voice-cloned calls and spoofed phone numbers. To create a good family password, avoid making it the same as any of your passwords and not sharing personal information that could be easily found by scammers. Consider anything you or your loved ones post online as data available to scammers, even if you keep all social media private.

Read the full story here

4. Blind people excluded from benefits of AI

The Royal Society for Blind Children's president, Tom Pey, has called for better design of AI tools and technologies, claiming that blind and partially sighted people are being excluded from the benefits of AI.

Pey claims that blind children are now distanced from their non-disabled peers, as they can experience games, alternative realities, and AI-driven visual types of technology. Pey, who lost his sight as a child, created the Waymap app, which offers step-by-step audio navigation instructions. He called on technology secretary Peter Kyle to formulate laws that support disabled people and direct big companies and startups to include disabled people. Research by the Royal National Institute of Blind People found that people with sight loss are less likely to use the internet daily, be digitally excluded, and own smartphones compared to the rest of the population.

Read the full story here

5. Qwen 2.5 Technical Report released