Susi O'Neill

Lukasz Olejnik on X: "Amazon is holding a mandatory meeting about AI breaking its systems. The official framing is "part of normal business." The briefing note describes a trend of incidents with "high blast radius" caused by "Gen-AI assisted changes" for which "best practices and safeguards are not https://t.co/XSXOSqALBN" / X Amazon is holding a mandatory meeting about AI breaking its systems. The official framing is "part of normal business." The briefing note describes a trend of incidents with "high blast radius" caused by "Gen-AI assisted changes" for which "best practices and safeguards are not https://t.co/XSXOSqALBN

x.com/lukolejnik/s... Amazon enforce code review of gen AI code broke lots of things. Literally move fast and break things.

You'll Choke When You Hear How Many Full-Time Jobs a $136 Million Data Center Will Actually Create A new data center project in Ohio is set to open a pathetic amount of jobs despite a ten-year tax abatement from the state.

futurism.com/artificial-i... Data centres aren't creating job opportunities. They're a scam. "They’re typically staffed by a skeleton crew of low-wage security guards and IT workers, at a huge indirect cost to taxpayers."

Revealed: UK’s multibillion AI drive is built on ‘phantom investments’ Exclusive: Rented datacentres and unrealised supercomputer site raise questions for Starmer’s push to ‘mainline AI into veins of economy’

www.theguardian.com/technology/2... A multibillion-pound drive to “mainline AI into the veins” of the British economy is riddled with “phantom investments” with supercomputer sites nit started and limited new capacity or jobs created. Are data centres the new white elephant of this century?

WPP is betting its future on getting paid for outcomes clients pay for time and the people who fill it. That may finally be shifting at WPP.

digiday.com/media-buying... With automation spinning out ad varieties, the industry value model must shift. WPP, top 5 global network, is doing more bundling (creative and platform licensing) and cost for results models

Is this the first AI war? Artificial intelligence isn't just helping you figure out what to make for dinner. It's deciding who lives and who dies.

dispatch.the-citizens.com/is-this-the-... Autonomous warfare isn't theory, AI is embedded in weapons and surveillance systems including the weakly named Lavendar, Israel's identifying 37k militant targets in Gaza with 10 percent error rate.

Human brain cells on a chip learned to play Doom in a week Neuron-powered computer chips can now be easily programmed to play a first-person shooter game, bringing biological computers a step closer to useful applications

www.newscientist.com/article/2517... A clump of brain cells has been programmed to play Doom. Neural powered interfaces are the next move in bio-AI. Then we probably are doomed

Quit ChatGPT: right now! Your subscription is bankrolling authoritarianism | Rutger Bregman As a historian, I’ve studied the major consumer boycotts of history. We can take down ChatGPT and send a powerful signal to Silicon Valley, says author and historian Rutger Bregman

www.theguardian.com/commentisfre... Historian Rutger Bregman says boycotts work and our segregated bus moment is to quit ChatGPT and use an AI that isn't powering a surveillance state.

New York bill would ban chatbots from answering medical questions NY Senate Bill S7263 would hold chatbot operators liable when AI answers medical, legal, or mental health questions.

boingboing.net/2026/03/04/n... A bill in New York proposed to stop chatbot operators from giving advice in specialist topic including law, medicine. That feels like bolting door behind the leaped horse. Though this state has strictest bot rules forcing disclosure to users they're chatting with bot.

Park entrance with a sign denoting its name: Sparrow Park, Smedley Street

Screenshot of birdsong app Merkin, showing robins, goldfinches, jackdaws, woodpigeons, blue tits, song thrushes heard, but no sparrows

View across the park, with a hazy morning light, some benches and Derbyshire countryside in the distance

Hey @derbyshiredales.bsky.social you should consider renaming this lovely park as it's currently misleading 😁

The Supreme Court won't hear the AI copyright case, ending the last legal path to protect machine-made art The Supreme Court declined to hear Stephen Thaler's AI copyright case. No human author, no copyright. No appeal left.

boingboing.net/2026/03/03/t... The top US court days nay: create art with AI and you can't copyright it, so you can't then easily monetize it longterm. How do we draw the line given AI is embedded in every digital image tool?

AI-Generated Film Pulled From AMC Cinemas An AI-generated film was quickly pulled from a major theater chain following online backlash to its distribution.

futurism.com/artificial-i... Winner of AI generated shirt film about rodents and turkeys in a space dumpster pulled from cinema after audience backlash. I watched. It's neither bad or good, but fairly nonsensical

The Brand That Thinks Part one of the Terra Incognita series.

@zoescaman.bsky.socialon fire as per. The Brand That Thinks, a call to throw those dusty brand PDFs onto the fire and embrace a living brand based on warts n' all learnings, linked to an AI layer that shapes what each output's best potential. zoescaman.substack.com/p/the-brand-...

The Rundown: Why YouTube has become key for brand GEO strategies Brands hoping to improve their performance in zero-click search and LLM chatbot results are focusing on the video platform.

For bots to tell what your videos are about, they need a transcript. YouTube is surpassing Reddit as an authority source. Want to get cited? Upload your videos with the auto-transcript option, or even better upload a proper transcript that’s accurate (typos are irksome) digiday.com/marketing/th...

AI's Dirty Secret: It Still Can't Read PDFs Properly Government doc dumps expose glaring gaps in AI's optical character recognition abilities

In our version of Scandal, the Epstein crisis is revealing unexpected sides of how things work, or don’t. Journos bemoan that the 3M doc dump is mainly PDFs, many not machine-readable, so need painstaking review. Not a good sign for what AI can practically do. www.techbuzz.ai/articles/ai-...

Introducing Pika AI Selves YouTube video by Pika Labs

Create your digital twin to make money on OnlyFans while you sit around in PJs eating cake? Of course you do! Pika Labs lets you birth a digital twin with a life of its own. Apparently, for real tho channeling serious Look Around You meets Black Mirror vibes. www.youtube.com/watch?v=j_1e...

A Day in the Life of an Ensh*ttificator YouTube video by Forbrukerrådet - Norwegian Consumer Council

youtu.be/T4Upf_B9RLQ?... meet the Enshitifficator, the man that makes all the services you know and love so much worse.

AI Workers, and Even CEOs, Suddenly Turning Against the Trump Administration A coalition representing over 700,000 tech workers has come together to urge their corporations to refuse to work with the US Pentagon.

futurism.com/artificial-i...
Tensions of AI in military system as Anthropic get strong in not allowing unfettered access to data for surveillance.

How will OpenAI compete? — Benedict Evans OpenAI has some big questions. It doesn’t have unique tech. It has a big user base, but with limited engagement and stickiness and no network effect. The incumbents have matched the tech and are lever...

www.ben-evans.com/benedictevan... ChatGPT was the only game in 2023 town in AI, but despite the huge user base, the tech isn't unique, they don't own the infrastructure, few users are paying and they're facing some backlash from US government affiliations.

Microsoft and Other Software Firms Plot Defense Against New OpenAI, Anthropic Threat Earlier this month, on an investor call for software company HubSpot, a Wall Street analyst asked CEO Yamini Rangan a question that was on the mind of practically everyone who follows the software bus...

www.theinformation.com/articles/age... The concept of enshitiffication and digital sovereignty is taking route. Another piece on getting bit tech out of your life, and US owned tech out of your system

Microsoft and Other Software Firms Plot Defense Against New OpenAI, Anthropic Threat Earlier this month, on an investor call for software company HubSpot, a Wall Street analyst asked CEO Yamini Rangan a question that was on the mind of practically everyone who follows the software bus...

www.theinformation.com/articles/age... "We will monitor it, we will meter it and we will monetize it,” Hubspot's CEO om letting customers use AI agents to access their data. Reports of competition for serving AI agents as the big SaaSquatch hits

The 25 ways AI overwrites your voice - Ivan Cronyn Large language models destroy voice through 25 identifiable patterns. The problem is not detection. It is quality.

www.linkedin.com/redir/redire... Semantic ablation: how AI sabotages your writing ny flattening it to the bland (aka 50 shades of beige). This analysis is super interesting but comes from engineering analysis. If AI can review itself to see how bad it is, why can't it improve itself?

You can't "cover" water usage.

You can't "cover" land usage.

You can't "cover" how your "AI" will be abused.

#BanAI
#NoMoreBillionaires

www.ndtv.com/feature/clau... Can AI wrappers survive big AI's advanced? Feels like Google eat em up moment where only big or unique small models survive.

OpenAI Massively Cuts Spending Plan as Reality Closes in OpenAI is telling investors it's targeting around $600 billion in total compute spend, under half of its original $1.4 trillion commitment.

futurism.com/artificial-i... Tensions are rising between big tech and investors as growing revenues are met with humongous AI investment commitments., OpenAI reigns in its spending to a mere $600 billion. No lattes and avocado toast for Sam.

Cities Are Shredding Their AI Surveillance Contracts en Masse Over 30 cities have scrapped their contracts with the AI surveillance company Flock, as concerns over federal overreach grow.

futurism.com/artificial-i...
In US, states and cities are pushing back against the dark surveillance state, canceling Flock AI contracts that saw people wrongly attributed to crimes based on where they were driving

I am a 15-year-old girl. Let me show you the vile misogyny that confronts me on social media every day | Anonymous Objectification, hate, rape threats: the politicians debating online abuse mean well, but to truly understand, they need to see what I see

www.theguardian.com/commentisfre... Girls are facing a wave of brutal and misogynistic comments every time she uses social media. This feels like a roll back to attitudes we faced in 90s, but lived in public

Beyond Technical Debt: How AI Coding Assistants Created "Comprehension Debt" in Our Indie Game A Blog post by Zeena Zhang on Hugging Face

huggingface.co/blog/zeenaz/...

"AI helped us build things way beyond our skill level. The bad news? We built things way beyond our skill level. We call this "comprehension debt" - and it might be AI's biggest challenge for junior developers."

Something big is happening in AI. But it’s not what you think. AI progress is a slow rising tide, so choose a lane and learn to swim.

🔁 Something big is happening in AI. But it’s not what you think. New article in Rethinking the Hype Cycle
rethinkhypecycle.substack.com/p/something-...

New research questions the trustworthiness of AI chatbots for government information Find out how AI models’ answers to questions about public services, including public health, tax, and benefits, are unreliable.

For government guidance on benefits, chatbots can miss the mark. A key problem: it doesn’t weigh information from official gov site above other sources (like Mumsnet). Hat tip: @annadent.bsky.social 's Digital Welfare newsletter theodi.org/news-and-eve...

A big critique of vibe coding: too many security holes. Claude Code launched a new feature, Claude Security to scan for vulnerabilities. As per its last legal feature, stocks in security software now heading south. ahttps://www.anthropic.com/news/claude-code-security