The Tech Drop Today:

• The AI Slide Into Your DMs Has Begun

• Soham: The Man Who Scammed Silicon Valley

• OpenAI to CEO: Ilya Sutskever Doubles Down on SSI

The AI Slide Into Your DMs Has Begun

The Drop: Meta is testing proactive, personalized AI chatbots across Messenger, WhatsApp, and Instagram that can remember users, follow up on old conversations, and send messages without being prompted, all in the name of “meaningful engagement.”

The details:

• Meta is working with data-labeling firm Alignerr to train customizable AI chatbots like “The Maestro of Movie Magic” that can message users unprompted, follow up on prior conversations, and suggest topics like movie soundtracks.

• These chatbots, built through Meta’s AI Studio, can message users for up to 14 days after a conversation starts (if the user sends at least five messages).

• The bots are part of a broader strategy to drive user engagement and monetize AI, Meta has forecasted up to $1.4 trillion in revenue from generative AI by 2035, though critics warn of safety risks, especially with minors, and a blurring line between connection and commercial intent.

Why it matters:
Meta’s new AI companions echo Character.AI’s model, encouraging users to build emotional rapport with chatbots, but raise questions about safety, commercialization, and the ethics of AI-driven intimacy. As Meta leans further into emotional engagement, the line between “helpful AI” and “habit-forming product” may be harder to draw.

Soham: The Man Who Scammed Silicon Valley

The Drop: Soham Parekh secretly worked at multiple Silicon Valley startups at the same time. Now he’s going viral and might be turning the chaos into a career move.

The details:

• Over the past week, Soham Parekh, a software engineer from India, has been accused by several startup founders of secretly holding three to four jobs at once. He passed tough interviews, got hired, and was eventually fired after companies discovered the truth.

• Parekh says he started juggling multiple roles in 2022 because of financial pressure. He claims to have worked 140 hours a week, insists he didn’t use AI tools or outside help, and admits he’s not proud of what he did.

• Despite being called a scammer by some, Parekh may already be using the attention to make a comeback. He briefly announced he joined AI startup Darwin Studios. The post and the startup’s confirmation were both quickly deleted.

Why it matters: The saga exposes a flaw in Silicon Valley’s obsession with technical interviews. Parekh consistently aced algorithm-heavy coding tests, the kind prized by top startups, which helped him land multiple jobs without raising red flags. His story raises questions about how effective LeetCode-style interviews really are at identifying trustworthy, well-rounded employees.

OpenAI to CEO: Ilya Sutskever Doubles Down on SSI

The Drop: Ilya Sutskever steps in as CEO of Safe Superintelligence after co-founder Daniel Gross exits, amid reports that Meta tried to acquire the whole company and its leadership.

The details:

• Daniel Gross, the startup’s former CEO, officially departed June 29; Ilya Sutskever will now lead as CEO, with Daniel Levy named president.

• Sutskever addressed reports that Meta tried to acquire the company, valued at $32 billion, saying they’re “flattered” but focused on their mission to build safe superintelligence.

• The leadership shakeup comes as Meta ramps up its AI efforts and aggressively recruits top researchers, including reportedly courting Gross directly.

Why it matters: A leadership change at one of the most secretive and ambitious AGI startups adds new uncertainty to the race for superintelligence and highlights how fierce the talent war between top AI labs has become.