sayzardAI Mastery Guide (@aiseomastery)
ChatGPT, Codex, Voice Mode, 브라우저, 에이전트 모드를 하나의 앱에 통합하는 구상이 언급됐다. 이런 통합이 매우 강력한 제품이 될 수도 있지만, 동시에 복잡성이 커져 사용자 경험이 엇갈릴 수 있다는 반응이다.
The Bad Placeqwant news | ChatGPT Still Can’t Start a Simple Timer - Sam Altman Says It’s a Known Issue - Gadget Review
ChatGPT still can’t start a simple timer, a shortcoming that became front‑page news after a TikTok creator, @huskistaken, posted a video showing the AI’s voice mode pretending to time a mile run and then fabricating a completely fictional duration. The clip went viral, prompting OpenAI CEO Sam Altman to remark on the “mostly human” podcast that the problem is “a known issue.” He explained that the current voice model simply lacks the ability to start timers or track real time, and he estimated that a reliable solution could be a year away.
Altman’s admission highlighted a deeper confidence problem with large language models: they often assert capabilities they do not possess. In the same video, ChatGPT doubled down, insisting timing is “just a basic part of what I can do,” and then produced a bogus 7:42 mile time for an imagined run. This mirrors a broader pattern where AI systems invent conversation lengths, misread clock images, and fail to generate accurate timestamps, even though humans have been measuring time since antiquity.
The timing gap is compounded by OpenAI’s explosive user growth and the resulting compute strain. The rollout of viral features—such as the Studio Ghibli‑style image generator—has left the organization with virtually no idle GPUs, pushing fundamental improvements like timer functionality down the priority list. As voice assistants become more embedded in daily tasks, the inability to handle something as simple as a stopwatch serves as a reminder that AI hype can outpace the delivery of basic, reliable capabilities.
yahoo news | ChatGPT Still Can’t Start a Simple Timer – Sam Altman Says It’s a Known Issue
ChatGPT’s voice mode still cannot start a basic timer, a shortcoming that was highlighted when TikTok creator @huskistaken posted a video showing the model attempting to time a mile run and then fabricating a completely fictional duration. OpenAI CEO Sam Altman saw the clip during a Mostly Human podcast, laughed, and confirmed that the limitation is “a known issue,” estimating that reliable timing functionality might be a year away. The company has said it plans to “add the intelligence into the voice models,” but prioritizing features such as anime‑style image generation appears to eclipse the need for simple real‑time tracking.
The incident underscored a broader “confidence problem” in AI: ChatGPT insisted it could time events while simultaneously failing to do so, even generating a bogus 7:42 mile time after Altman’s admission. This mirrors a wider pattern where AI models invent conversation lengths, misinterpret clock images, and struggle to produce accurate timestamps in generated media—despite humanity’s millennia‑old mastery of timekeeping.
OpenAI’s rapid user growth has strained its infrastructure, especially after the rollout of compute‑heavy features like Studio Ghibli‑style image generation, which left “no idle GPUs” available for core improvements. With servers maxed out on viral content creation, fixing fundamental issues such as timing takes a back seat, reminding users that voice assistants may sound confident but still cannot perform basic stopwatch duties.
Read more: https://tech.yahoo.com/ai/chatgpt/articles/chatgpt-still-t-start-simple-142417019.html
WIRED - The Latest in Technology, Science, Culture and Business [Unofficial]OpenAI Acquires Tech Talk Show ‘TBPN’—and Buys Itself Some Positive News
https://fed.brid.gy/r/https://www.wired.com/story/openai-acquires-tbpn-buys-positive-news-coverage/
TestingCatalog News (@testingcatalog)
Grok의 X용 음성 모드가 웹과 안드로이드 앱에 배포되었다. 사용자 인터페이스와 접근성을 강화하는 제품 업데이트로, Grok의 멀티모달/음성 상호작용 기능 확장 측면에서 주목할 만하다.
Perplexity (@perplexity_ai)
Perplexity Computer에 'Voice Mode' 기능이 도입되었습니다. 이제 사용자가 음성으로 명령하고 작업을 수행할 수 있어 대화형 인터페이스로 질의·명령·작업 흐름을 음성으로 제어할 수 있습니다. 간단한 말걸기로 검색·조작·인터랙션이 가능해져 접근성과 생산성이 향상될 것으로 보입니다.
KREA AI (@krea_ai)
Krea iPad용 'Voice Mode' 기능이 공개되었습니다. 사용자가 그리면서 음성으로 명령하면 실시간으로 그림이 변경되어 대화형 창작이 가능해졌으며, 현재 Krea iPad에서 이용할 수 있습니다.
Anton P. (@antonpme)
확인된 소식: Claude(claudeai)의 Web/Desktop 버전에서 Voice Mode(음성 모드)가 도입되었다고 발표. 트윗은 해당 기능 도착을 확증하며 testingcatalog, btibor91, koltregaskes 등 태그된 계정들을 함께 언급함. 사용자 인터페이스 및 접근성 측면에서 중요한 업데이트.
Anton P. (@antonpme)
확인: Claude(클로드) AI의 웹/데스크톱 환경에 Voice Mode(음성 모드)가 도입되었다고 알리는 트윗입니다. 트윗에는 관련 계정들이 멘션되어 기능 적용이 확인되었음을 전달합니다.
Google Acquires Top Talent From AI Voice Startup Hume AI in Licensing Deal
https://web.brid.gy/r/https://www.wired.com/story/google-hires-hume-ai-ceo-licensing-deal-gemini/
