Welcome again to a different entry of my weekly AI Diaries. Each time I sit down to write down these updates, I really feel like I’m attempting to catch water with my naked fingers—the tempo of innovation in synthetic intelligence is simply that relentless.
We used to take a look at AI by means of extremely rose-tinted glasses. We talked endlessly about how it could treatment illnesses, revolutionize clear power, and act as a universally constructive power for humanity. We even noticed large tech conglomerates signal manifestos swearing they might by no means weaponize these fashions. Nicely, after wanting on the information crossing my desk this week, it’s protected to say these guarantees have formally been forgotten.
Let me stroll you thru what occurred during the last seven days, as a result of we simply crossed a number of main thresholds which are going to basically change how we work together with know-how.
The “Darkish Manifesto” and the Militarization of AI
The most important shift I seen this week wasn’t a brand new piece of software program; it was a shift in ideology. The boundaries of how AI can be utilized are being aggressively redrawn, and virtually talking, every part is now on the desk.
The clearest indicator of this pivot got here from Palantir. They simply printed a 22-point doc titled the “Technological Republic” manifesto. Studying by means of it gave me chills—social media is already calling it the “darkish manifesto,” and for good cause. It primarily normalizes the usage of AI as a weapon. Palantir, whose affect within the US protection sector is quickly increasing, used this doc to publicly argue that using AI for deadly army functions isn’t solely acceptable however obligatory. It’s an open declaration of what all of us sort of suspected was occurring behind closed doorways.
They usually aren’t appearing alone. Two large developments dropped proper alongside Palantir’s manifesto:
Anthropic’s Reversal: Anthropic beforehand acknowledged their most superior mannequin, Mythos, might be extremely harmful within the improper fingers. Now? It seems they’re exploring opening federal businesses’ entry to it.Google and the Pentagon: Reviews broke that Google is at the moment in talks with the Pentagon to create a framework that may enable Gemini to function in extremely labeled, safe environments.
The AI arms race is not a sci-fi idea; it’s our present actuality.
Google Chrome Will get a Gemini Improve: Welcome to the “Expertise” Period
Shifting away from the heavy stuff, Google dropped a genuinely improbable replace for these of us attempting to hurry up our each day workflows. They’re deeply integrating Gemini into the Chrome browser with a brand new characteristic they name “Expertise.”
Right here is the way it works: As a substitute of typing out the identical advanced prompts over and over, now you can save them as a “Talent.” Let’s say you will have a selected means you want AI to summarize lengthy articles or format information tables. You save that immediate, and the following time you want it, you simply kind a “/” within the Gemini interface or click on a plus button, and your customized Talent runs immediately in your lively tab.
What I actually love about that is the ecosystem integration. If you’re signed into your Google account on desktop, your saved Expertise sync seamlessly throughout your gadgets. It will probably even pull information from a number of open tabs concurrently. It’s a large time-saver.
Microsoft’s Panic Button: “Copilot Code Crimson”
It looks like Microsoft is feeling the warmth. Regardless of pushing Copilot into principally each product they personal, they aren’t seeing the consumer adoption or efficiency outcomes they wished. To stop falling behind within the AI race, they’ve initiated a large inside pivot formally dubbed “Copilot code purple.”
The objective is to drastically overhaul the consumer expertise, however the technique is what fascinates me. Simply days after this “code purple” leaked, we came upon Microsoft is testing methods to show Copilot right into a 24/7 autonomous digital assistant.
By integrating AI brokers just like the OpenClaw structure into the Copilot ecosystem, Microsoft desires this software to cease ready on your prompts. They need it operating within the background, independently executing duties and managing your digital life across the clock. If they’ll pull this off with out it turning into an intrusive mess, it may change every part about how we use Home windows.
Robots Are Lastly Getting “Sensible”: Gemini Robotics-ER 1.6
Should you learn my articles recurrently, I’ve a tender spot for robotics. This week, Google DeepMind blew my thoughts with their new Gemini Robotics-ER 1.6 mannequin.
Till now, most robots simply blindly adopted pre-programmed scripts. DeepMind’s new mannequin offers machines “embodied reasoning.” This implies the robotic doesn’t simply see its setting; it really understands the context. It will probably analyze visible information, plan a multi-step job on the fly, and self-evaluate whether or not it accomplished the job accurately.
Boston Dynamics wasted completely no time placing this into motion. They geared up their well-known robotic canine, Spot, with the ER 1.6 mannequin, and the demo video is wild. Spot is proven a handwritten to-do checklist and simply… doing the chores. I watched a robotic canine set up scattered sneakers, choose up bins, and actively type laundry right into a basket. In a single scene, it even grabbed a leash to take an actual canine for a stroll based mostly totally on pure language instructions. We’re watching the transition from senseless machines to reasoning bodily entities in real-time.
The Posthumous Return of Val Kilmer
Lastly, we’ve to speak concerning the moral grey space Hollywood simply sprinted into. The legendary actor Val Kilmer, who sadly handed away final yr after his lengthy battle with throat most cancers, is “starring” in a brand new movie referred to as As Deep because the Grave.
Earlier than his demise, Kilmer had agreed to play the function of an American priest within the movie, however his well being prevented him from ever taking pictures a single scene. Now, with the specific permission of his daughter, the studio has used AI to utterly recreate Kilmer’s likeness and voice to play out all the function.
Even with the household’s blessing, this has triggered a large uproar within the leisure business. Whereas the know-how is undeniably spectacular, it leaves us asking deep questions on legacy, artwork, and the commercialization of digital ghosts.
Speedy Fireplace: New AI Instruments & Business Bites
A variety of smaller, however equally vital, updates flew below the radar this week. Right here is my fast breakdown of the remainder of the information:
New Device Drops:
Claude Opus 4.7: Anthropic launched what they’re calling probably the most highly effective mannequin at the moment obtainable to most people.Claude Design: Powered by Opus 4.7, this new suite combines prototyping, slide deck creation, and visible technology. It’s particularly constructed for individuals with zero design background.Blissful Oyster & Lyra 2: Each Alibaba (Blissful Oyster) and Nvidia (Lyra 2) launched large instruments that allow customers generate absolutely interactive, navigable 3D worlds utilizing AI.Gemini 3.1 TTS: Google dropped an replace to their Textual content-to-Speech mannequin, and albeit, it’s one of the vital natural-sounding vocal technology fashions I’ve heard thus far.
Business Shorts:
China’s Grasp Plan: China is rolling out a large nationwide initiative to combine AI instantly into their schooling system, from elementary colleges all the best way to lifelong grownup studying. They’re aggressively constructing a digital-first workforce.Meta & Broadcom: The 2 giants signed a long-term deal (stretching to 2029) to co-develop customized AI silicon chips.Opera Browser Connector: Opera One and GX browsers simply obtained an replace that permits ChatGPT and Claude to natively learn open tabs and work together instantly with the browser interface.Fixing the Unsolvable: A Chinese language AI mannequin independently solved a fancy math downside proposed by American mathematician Dan Anderson in 2014—utterly with out human intervention.Nvidia Ising: Nvidia revealed the primary AI mannequin able to breaking quantum limitations. “Ising” accelerates calibration and error correction in quantum computer systems, boosting efficiency by 2.5x and accuracy by 3x.Unconscious Bias: A captivating new examine printed in Nature proved that when AIs generate coaching information for different fashions, they cross down hidden biases and prejudices on a “unconscious” degree.Native Mac Gemini: Google formally launched a standalone, native Gemini app for macOS.Vercel Hack: The cloud deployment platform Vercel confirmed a current cyberattack. The scary half? The breach originated totally from the hijacking of a third-party AI software.
Over to you: I wish to hear your tackle the Val Kilmer state of affairs. If a household offers their blessing, do you assume it’s acceptable for studios to resurrect deceased actors utilizing AI for model new films, or ought to a performer’s filmography finish after they cross away? Drop your ideas within the feedback, as a result of I’m genuinely torn on this one!
