In partnership with

Zuckerberg’s Grand Vision: Most of Your Friends Will Be AI

Key Takeaway: Mark Zuckerberg envisions a future where AI companions outnumber human friends, stepping in as therapists, customer-service agents, and social connectors to meet people’s unmet relationship needs.

More Insights:

• Zuckerberg warns the average American has fewer than three friends but craves around 15, positioning AI “friends” as a solution to widespread loneliness.

• He predicts everyone will use AI therapists, filling gaps where human mental-health care is inaccessible or unaffordable.

• Meta is already integrating “Meta AI” into Instagram, Facebook, a standalone app, and Ray-Ban smart glasses—drawing nearly a billion monthly users to its chatbots.

• Critics argue AI companions risk deepening social isolation and lack the genuine empathy of human relationships, while proponents see them as better than no support at all.

Why it matters: As AI permeates our social fabric, we must grapple with whether algorithmic companionship can truly fulfill human needs or if it will reshape—and potentially erode—the essence of genuine human connection.

Related: Zuckerberg’s new Meta AI app gets personal in a very creepy way

The Future of Voice AI Is Here

Discover why forward-thinking enterprises are rapidly adopting Voice AI Agents. This guide breaks down the $47.5B market shift, highlights emerging trends, and offers practical steps for successful implementation.

Learn how leading teams are using Voice AI to boost efficiency, elevate customer experience, and start delivering measurable results—in as little as 3 weeks.

Download the free guide

Open source project curl is sick of users submitting “AI slop” vulnerabilities

Key Takeaway: Daniel Stenberg, lead of the 25-year-old curl project, is cracking down on the flood of AI-generated, low-quality vulnerability reports, warning that they’re effectively “DDoSing” maintainers’ time.

More Insights:

• AI-slop detection: Reports often feature perfect English, bullet points, and even pasted AI prompts—telltale signs of machine-generated noise.

• Bogus HTTP/3 exploit: A recent submission claimed a novel stream-dependency attack leading to remote code execution, but cited non-existent functions and failed to apply to the current codebase.

• HackerOne stance: While Stenberg deems AI reports useless, HackerOne insists AI can aid researchers if used responsibly and vows to treat hallucinated submissions as spam.

• Community proposals: Suggestions include requiring disclosure of AI use, charging a review bond, or building new infrastructure to filter out “slop” before it reaches maintainers.

Why it matters: As AI tools democratize technical writing, they also blur the line between genuine research and automated fluff—forcing open-source projects to rethink how we verify, reward, and trust security contributions in an era of omnipresent LLMs.

U.S. AI czar David Sacks projects that AI will go 1,000,000x over the next four years

Key Takeaway: David Sacks argues that simultaneous exponential advances in algorithms, chip design, and compute capacity will multiply AI’s capabilities by a millionfold over the next four years.

More Insights:

• Models evolve rapidly: LLMs are improving ~3–4× per year, transitioning from chatbots to reasoning engines and soon to autonomous agents.

• Hardware leaps: Each new GPU generation delivers ~3–4× better performance, and innovations like NVL72 racks amplify datacenter throughput.

• Compute scale-up: Major players have grown GPU fleets from hundreds of thousands to potentially millions, fueling ever-larger training runs.

• Compound effect: At 10× every two years per dimension, four years of progress yields 100× per area—and 100׳ = 1,000,000× overall.

Why it matters: This hyper-accelerated trajectory will drive dramatic cost cuts, unlock capabilities we can’t yet imagine, and embed AI into every facet of the economy—transforming industries and society at warp speed.

Other stuff

• Connecting GitHub to ChatGPT deep research

• Google rolls out AI tools to protect Chrome users against scams

• IRS hopes to replace fired enforcement workers with AI

• 18-Year-Old U.S. Teen Develops AI to Reveal 1.5 Million New Variable Space Objects

• OpenAI and the FDA Are Holding Talks About Using AI In Drug Evaluation

• Apple is looking to add AI search engines to Safari

• For AI Startups, A 7-Day Work Week Isn’t Enough

• Netflix introduced a new AI-powered interface

• Figma adds AI across its design suite

• OpenAI stays non-profit

Hope you enjoyed today's newsletter

⚡️ Join over 300,000 people using the Superpower ChatGPT extension on Chrome and Firefox.