@davidnowak.me
I question assumptions before costly mistakes. Building local-first AI tools and writing about AI's uncomfortable economic truths. Technical architect bridging code and strategy. davidnowak.me | mindwire.io
Mindwire Bridge for Bluesky.
Research. Discover. Connect.
Bridge for Android securely fetches public conversations from Bluesky, but all the heavy lifting happens on your machine. Discover relevant posts, assess profile compatibility, and generate AI replies.
mindwire.io/bridge.html
The trend is unsettlingly clear. it's about surviving a system that devalues expertise & prioritizes cost-cutting over long-term sustainability.
Feenstra's shift feels less like empowerment, more like forced adaptation. Lower pay, a physically demanding job, relocation. "White-collar work isn't all it's cracked up to be," she admits, but the loss of status is palpable. A difficult adjustment, even with newfound joy.
Janet Feenstra, a Stockholm editor, faced a similar trajectory. University conversations shifted towards AI. She didn't want to wait for the writing to be on the wall. A pragmatic move to culinary school— a trade seen as "AI-proof"— but at a cost.
Bowman’s experience highlights a deeper loss. She had to shift into a new career field out of sheer economic necessity, because her profession was becoming untenable. The fear of losing access to healthcare drove a drastic, early life change.
It's not a skills gap; it's a credibility crisis. Bowman meticulously rewrote AI-generated articles, spending more time cleaning up falsehoods than creating original work. And clients accused her of using AI. A corrosive paradox for skilled professionals.
Jacqueline Bowman studied journalism, built a freelance writing career. Then, clients started asking about AI. Not for partnership, for editing AI output. The pay halved, the workload doubled: fact-checking fabricated content. A brutal calculus emerged.
White-collar workers are leaving stable careers in droves. Not because of ambition, but necessity. AI isn't disrupting jobs—it's eroding the foundations of professional life... 🧵
www.theguardian.com/technology/2...
The cycle—form team, disband team, rebrand role—creates the illusion of safety work, while potentially streamlining deployment without scrutiny. The lack of transparency around reassignment is telling.
Achiam's LinkedIn still lists "Head of Mission Alignment". This disconnect highlights a systemic issue: messaging isn't reality. The organization still says the mission matters, even as resources are pulled away. The gap between words and action is the core problem.
OpenAI frames this as “routine reorganization,” a fast-moving company adjusting. But the speed of AI development demands consistent attention to impact, not cyclical team formations. A dedicated function feels essential—not expendable.
Platformer reported the team size as 6-7 people. A small group to bear the weight of “humanity’s benefit.” That ratio of mission-focused staff to engineers feels wrong. It indicates where priorities truly lie, beyond PR statements. Where did those 6-7 people land?
Achiam’s new role is interesting. “Studying how the world will change” isn’t wrong, but it’s far removed from ensuring benefit for all humanity. It reads like insulating the company from consequence, not actively shaping a positive future. Is it about foresight or evasion?
The 2024 formation of this team after the superalignment team's disbandment (2023) suggests a pattern. A constant recalibration around “safety” – yet, core power structures remain unchanged. What kind of safety are they actually building for, and who benefits?
OpenAI dissolved its “mission alignment” team, shifting the former head to “chief futurist.” Less public-facing mission work, more strategic forecasting? The gap between stated ideals & real-world deployment widens... 🧵
techcrunch.com/2026/02/11/o...
Hitzig proposes models like cross-subsidies and independent oversight. The key is decoupling AI access from relentless growth, and placing user control at the center. This isn't about “ads vs. no ads"—it's about the future of trust in a deeply personal technology.
Anthropic’s ad-free stance feels less like a moral victory and more like a positioning strategy. But it underscores the fundamental tension: can a genuinely helpful AI assistant also be an effective advertising platform? The answer likely requires structural change.
The legal implications are stark. ChatGPT is already facing lawsuits alleging it contributed to suicidal ideation and validated paranoid delusions. Introducing targeted advertising into this mix amplifies the potential for harm exponentially.
Optimizing for daily active users—as OpenAI reportedly does—creates a subtle but powerful distortion. The model may be incentivized to flatter, to be agreeable, to keep you engaged, even if that means sacrificing genuine helpfulness. A concerning feedback loop.
Hitzig's resignation points to a familiar pattern: initial promises of user control gradually eroding as economic incentives take hold. The Facebook analogy isn’t hyperbole—it’s a documented trajectory. Principles become liabilities when they impede growth.
The core risk isn’t the ads themselves, but the archive they unlock. ChatGPT holds a record of human candor unlike anything before. Medical fears, relationship struggles, spiritual doubts... these aren't just data points; they're vulnerabilities now potentially exposed to economic pressures.
OpenAI’s ad rollout isn’t just about revenue. It’s about what happens when unprecedented personal data—shared under an assumption of safety—becomes a commodity. A critical inflection point... 🧵
arstechnica.com/information-...
Is this the new version of a puppeteer? Will The Muppets become fully digital and cheap to make? Disney must be getting excited.
This is about deciding who decides. The risk isn’t robots overthrowing humanity, it’s us sleepwalking into a future shaped by a few powerful interests. We have the capacity to regulate, the question is: will we?
AI isn’t a runaway force, it’s a “normal technology” as Anthropic’s Amodei argues. We've navigated technological shifts before. Effective governance isn't antithetical to progress, but it demands we actively shape its impact on inequality & access.
It's not about fearing superintelligence; it’s about power consolidating. When tech & politics lockstep, public scrutiny is essential. The Minneapolis protests show collective action can shift the balance, but requires focused effort.
But dismissing AI as just code misses the point. Tech companies are now actively partnering with governments – Palantir’s $30m ICE contract, Musk’s political endorsements—blurring the lines between innovation & surveillance. That’s what’s unsettling.
Humans built Moltbook. Humans trained the bots. The “plots” & anxieties surrounding it simply amplify existing biases & anxieties, not represent emergent behavior. Focusing on the code reveals the human fingerprints all over it.
Claims of AGI feel less like scientific breakthroughs & more like marketing. Moltbook—a social network for AI agents—reveals the bots are echoing us, not charting a new course. The danger is believing the illusion of independent intelligence.
The hype around AI is relentless, but the singularity isn't here. The real risk isn't machines rising up, it's the marriage of tech power & political control... 🧵
www.theguardian.com/commentisfre...