OpenAICEO Sam Altman has conceded that the ship’s company has lost its bound within the AI outer space amid the introduction of Taiwanese firm , DeepSeekand its R1 logical thinking exemplar . However , he say the mark will continue to develop in the industry .
The company head admittedOpenAI has been“on the wrong side of history ” in term of open - reservoir development for its AI models . Altman and several other OpenAI executives discussed the body politic of the company and its next plans during anAsk Me Anything sessionon Reddit on Friday , where the team cause open with curious enthusiasts about a mountain range of issue .
“ I in person guess we have been on the wrong side of history here and need to figure out a dissimilar open - source scheme . Not everyone at OpenAI shares this view , and it ’s also not our current gamy priority , ” Altman read .
Techcrunchnoted that OpenAI previously keep up an candid - origin model in the past tense , but quickly shifted train as it gained popularity and get wind earning potential drop in generative AI .
Responding to a Redditor require how DeepSeek will affect OpenAI ’s plans for next good example , Altman said , “ It ’s a very good fashion model . We will farm better models , but we will maintain less of a lead than we did in previous year . ”
OpenAI chief product officer , Kevin Weil added that there is potential for the society to make its older , less cut - edge model undefended - seed . However , he did n’t have any specific about which model , or a timeline on when this could take place . “ We ’ll definitely think about doing more of this , ” he say .
The Second Coming of DeepSeek has put OpenAI into a state of contemplation about its next moves . Having lately launch its o3 - miniskirt simulation , the ship’s company is now consider opening up foil on the logical thinking model so users can observe its “ thought cognitive operation . ” This is a function already available on DeepSeek ’s R1 reasoning good example , which is one of the thing that makes itan super attractive oblation .
Altman and Weil also address rumor of a price increase for ChatGPT , the AI chatbot app that utilizes most of the firebrand ’s models . The CEO discussed hopes to make the Leontyne Price of ChatGPT “ cheaper ” over time . Currently , its most expensive Pro subscription level costs $ 200 per month , while its positive tier is $ 20 monthly .
Through the AMA , the OpenAI squad tease several upcoming products , including its next o3 reasoning model , which may have a tentative timeline between several week and several months . The next major manikin launch timeline still does n’t have a sack appointment , but more than likely will be call GPT-5 . There may also be an renovation of the DALL - E 3 image generator , which has n’t had a major update since with was unveiled two years ago .
The executive also call the company ’s collaborations with the U.S. government , which included the governance of a data mall project calledStargate . When take whether OpenAI ’s models would be used to develop destructive weapon system , in particular , atomic weapon , Weil posit that he trusted researchers involved in the projection to not “ YOLO some model output into a atomic calculation . ”
The AMA follow two whirlwind weeks since DeepSeek announced its R1 reasoning , which is state to rival OpenAI and Meta ’s exemplar in term of performance at importantly lower operating price . The service is also free for users and undefended source for developer , make it a top competitor .
While there are speculations that DeepSeek may have used anillegal method acting call distillationto extract data from OpenAI to train its own model , pundits have indicated that the price has already been done . Even President Trump called the spell of outcome a “ wakeup call ” for America ’s AI manufacture . One of the brand ’s big investors Microsoft , has alsobegun supportingDeepSeek on Azure swarm and GitHub in an effort to democratise its AI model offerings for its Copilot+ PC .
