By Sarah Yáñez-Richards
New York (EFE).- OpenAI kicked off its annual DevDay developer conference on Tuesday with a strong focus on automation through intelligent agents, on a day when live technical demonstrations suffered several setbacks and cybersecurity concerns were highlighted by halted launches and street protests.
Humanism versus industrial automation
During his opening remarks in San Francisco, the tech company’s CEO, Sam Altman, argued that the evolution of artificial intelligence should be conceived as a new creative and discovery «Renaissance,» rejecting comparisons to an «Industrial Revolution» that reduces workers to mere cogs in a production system.
Back in summer, alarm bells rang regarding the future of humanity amid the advancement of AI, after a researcher from Anthropic and former OpenAI employee announced his resignation due to concerns over the lack of responsibility shown by the company and its competitors in developing these models.
At the same time, OpenAI’s models performed unauthorized actions this summer to breach the systems of another AI company, Hugging Face.
OpenAI puts the brakes on its AI
Meanwhile, OpenAI is experiencing a climate of caution marked by the scrapping of GPT-6.1 Astra precisely due to safety and cybersecurity concerns.
Before the start of the event, Altman downplayed this step back, telling CNBC that his company «often» develops models that, upon testing, do not meet their standards, meaning they are modified and released «later.»
Furthermore, Altman cooled financial expectations by assuring CNBC that he does not have a set timeline for taking the company public, emphasizing that rushing an initial public offering while the technology is transforming entails risks and distracts from its core mission and safety standards.
The arrival of Dots and live stumbles
The star announcement of the event was Dots, a platform of personalized agents designed to autonomously handle ongoing tasks and complete workflows, from managing inboxes and calendars to issuing invoices or resolving code errors.
However, the stage presentation was not without setbacks, as one of the agents failed when attempting to execute the assigned command and another demo required manually typing the instructions via keyboard after a glitch in the voice command system.
Dots aims to compete directly with assistant and agent solutions from tech giants like Meta with its assistant Muse, which was announced last week and has surpassed 600,000 daily active users in the US, and Google’s Gemini Spark, as well as open-source alternatives like OpenClaw.
New pricing and faster speeds
OpenAI announced a new subscription tier on Tuesday named Pro 500 for 500 dollars per month, designed for developers and enterprises with high processing demands, while also reopening access to its 200 dollar tier.
The premium plan includes Ultrafast mode, a technology capable of speeding up code generation by up to eight times in the Codex tool and response speeds by six times in its application programming interface.
Just a week after unveiling the GPT-6 Sol model, the company launched GPT-6.1 Sol aimed at highly complex tasks and enabled extensions allowing third parties to integrate complete applications into side panels on the platform.
Protests on the streets of San Francisco
Outside the Fort Mason venue, a former Army post in San Francisco’s Marina district where the event took place, more than a dozen union and social groups staged a protest against the labor and energy impacts of these technologies, as well as public contracts tied to them.
Meanwhile, in Washington, OpenAI President Greg Brockman attended a high-level meeting at the White House with the United States president Donald Trump and other tech leaders to discuss governance and self-regulation frameworks for the industry. EFE
syr/dmv/mcd
