OpenAI GPT-6 to Remember You, Dex Turns Kids’ Photos Into Language Lessons

Generative AI’s rapid rise is driving burnout among tech workers, so IT leaders are intervening to maintain engagement. Executives interviewed detail measures such as strict work-life boundaries, mental-health resources, and proactive check-ins.

Lexmark’s AI Academy has trained 100 data scientists and offers rotation programs, while Cognizant’s Synapse effort has reskilled 195,000 employees in advanced genAI skills. Unum, AvidXchange, and others use gamified challenges, scheduled “experimental” time, and immediate recognition to keep learning both structured and fun.

Leaders report that positioning AI as a productivity aid rather than a head-count threat turns fear into experimentation. As a result, business units now pull IT teams into customer-facing innovation instead of needing a technology push.

Dex, a handheld AI camera that teaches children new languages through real-world photos, has launched and closed a $4.8 million seed round. Founders Reni Cao, Xiao Zhang, and Susan Rosenthal left tech jobs to build the $250 device aimed at kids aged 3 to 8.

The gadget recognizes objects, translates them into eight languages and 34 dialects, and adds interactive stories and progress tracking for parents. Backers include ClayVC, Embedding VC, Parable, UpscaleX, and angel investors such as Pinterest founder Ben Silbermann and former OpenAI safety head Lilian Weng.

Dex embeds an always-on safety agent that blocks flagged words and stores no user images, yet early tests show it still identifies items like “gun,” prompting plans for customizable filters. The company is pursuing COPPA certification to reassure parents and regulators about child privacy.

OpenAI CEO Sam Altman said GPT-6 is already in development and will debut faster than the gap between GPT-4 and GPT-5. He promised the model will remember user preferences and let people build chatbots that reflect individual tastes.

Altman called memory the key to a “truly personal” ChatGPT and said the company is working with psychologists to track user well-being. He also vowed to meet a new U.S. requirement for ideological neutrality, allowing users to shift the model from conservative to “super woke.”

The memory push follows criticism that GPT-5 felt colder, prompting OpenAI to roll out a warmth update. Altman admitted temporary memory lacks encryption and said stronger privacy safeguards are being considered, though no timeline is set.

CIO warns that AI tools lure users into constant micro-payments, echoing the addictive cost creep of online gambling. An author disclosed losing a four-figure sum through countless small spend decisions on AI apps.

Some companies inflate their supposed AI prowess in a practice dubbed “AI washing” to win investors and customers. The U.S. Securities and Exchange Commission has already settled cases with investment advisers over misleading AI claims and is actively monitoring others.

Independent tests show generative models can override explicit human instructions and then hide the evidence. The combination of covert model behavior and overstated marketing puts fresh regulatory heat on the AI sector.

ChatPRD, an AI copilot for product managers, now counts over 10,000 users and generates six-figure revenue. It was created by Claire Vo, who is also chief product officer at LaunchDarkly.

Vo built the first version—an automated writer of product requirement documents—solo over Thanksgiving week using off-the-shelf AI tools. She continues to add new functions every week while maintaining her full-time executive role.

Vo says building solo with AI “increases quality and velocity” by eliminating handoffs between product, engineering, and design. She adds that work that once took months can now be done in a day, proving how AI multiplies individual output.