Cognitive warmup. Every action has an equal and opposite reaction. If you keep shouting the same thing every day, that AI will replace humans in the workplace, someone at some point will lose their patience.
I wouldn’t for a moment condone physical violence, yet to a degree, the reaction of a 20-year old who threw a Molotov cocktail at OpenAI CEO Sam Altman’s home should also be weighed from the perspective of joblessness and a bleaker-than-bleak future as a result. I’ve often asked what AI companies and AI bros are trying to achieve by making humans redundant in workplaces. How do they intend to deal with millions of jobless humans walking the face of the earth? Perhaps it’ll then become a civil or law and order issue, not their problem?
Oh, don’t give me that “alternative careers will come up” crap (no AI bro has ever been able to list one alternate career to look forward to; they just talk in circles). And for Altman to blame the violence on what he feels is somewhat critical coverage of OpenAI and his pursuits, simply lays bare the mindset and priorities of the man. But then again, it’s a mindset we already know of. He has once drawn parallels between the food a human eats in their lifetime versus the energy it takes to train an AI model (and the AI bro concluded that AI is better).
This whole AI enthusiasm underlined by well-documented circular financing is a money grab as far as it’ll go. AI will remain, but it’ll remain a tool for humans. Not replace humans. Unless we’re willing to be replaced. I, for one, am not surrendering reasoning and cognition to an AI chatbot. Not in this lifetime.
Every action has an equal and opposite reaction
Apart from rendering millions of humans jobless, another big AI bro fascination over the past year has been data centre investments—billions upon billions of dollars spent and vast amount of energy and water guzzled for AI slop on social media. It’s good to see the pushback has started in the US with the Energy Information Administration expected to implement a nationwide policy that would require data centres to declare how much power they consume.
This would be the first of its kind initiative to collect this information from data centres. One would’ve assumed it’s basic disclosure but that hasn’t been the case so far.
The State Legislature of Maine has passed a law to ban large data centres. Any project that calls itself a large data centre has always led to an increase in energy costs in that geography. There are more US states and cities which are weighing similar legislation. For now, Maine puts a moratorium on construction till late 2027, and in the meantime, a council will be formed that will evaluate any data centre requests weighed against potential cost of living for Maine’s citizens. By the way, here’s a neat crowd-sourced Data Center Proposal Tracker that you might like to browse for a bit.
Mac users! A useful Gemini app
A native Gemini desktop experience is finally arriving on macOS. The native desktop experience means a user can share anything with Gemini or invoke the AI assistant without having to switch windows or apps.
The shortcut is Option+Space to invoke a floating chat bar—and everything will be accessible here exactly how Gemini is on your smartphone, or within the web browser, including uploading files or photos from Google Drive for queries or image generation.
A user will need to give explicit permission to Google Gemini to access the system’s information before sharing a window—Mac’s system settings > Privacy and security and enable Accessibility for Gemini. That will be needed if the intent is to share what’s on the screen for context to a query. The Gemini app for macOS is available globally now, and Gemini access tiers will depend on the subscription plan.
A new Claude, that claims to be better
Anthropic says this is their most powerful generally available model till date, but “its cyber capabilities are not as advanced as those of Mythos Preview”. No surprise that Anthropic is referencing the Mythos claims since a lot was said about it securing the world’s software and whatnot. Anyway, Claude Opus 4.7 is a “notable improvement on Opus 4.6 in advanced software engineering, with particular gains on the most difficult tasks”. You don’t say!
Anthropic says Opus 4.7 is more tasteful and creative when completing professional tasks. It claims higher benchmark scores than Google Gemini 3.1 Pro and ChatGPT 5.4 in agentic coding, multidisciplinary reasoning, and agentic financial analysis. The pricing remains the same as Opus 4.6, at $5 per million input tokens and $25 per million output tokens.
Neural Dispatch is your weekly guide to the rapidly evolving landscape ofAI. Want this newsletter delivered in your inbox? Subscribe here.