#Ai
53 notas · última:
How AI does, and does not, change the way I do math
Rachel Webb argues that LLMs will change how she executes math research, but not what makes it interesting. She frames math as both personal enjoyment and a community practice.
Linus Torvalds: 90% of AI marketing is hype so 'I ignore it'
Linus Torvalds says most AI marketing is hype, calls the industry 90 percent marketing and 10 percent reality, and says he will ignore GenAI for now until real workloads show up.
8 Predictions for 2026. What comes next in AI?
The page predicts 2026 will bring generative UI, edge-based personal agents, smarter home assistants, AI-labeled content, and biometric verification as AI systems spread into daily life.
The End of Programming as We Know It
The article argues that AI will not replace programmers, but it will change programming into higher-level work, while creating new demand, new skills, and new roles around AI systems.
I am tired of AI
The author says AI is useful in some cases, but argues that marketing hype, AI-generated test tools, AI-written conference proposals, and dull synthetic content have become exhausting and overused.
Power in the Age of Intelligence
The essay argues that AI and other new technologies increase concentration: the right question is not how software keeps moats, but which companies own scarce industry assets, break constraints, and expand from the high ground.
Among the Agents
The author argues that coding agents have become capable enough to do substantial real work, but the real shift is using them well. He sketches practical limits, risks, and likely effects on software, data, jobs, policy, and growth.
AI #156 Part 1: They Do Mean The Effect On Jobs
The post argues that recent AI progress is already showing up in productivity, jobs, legal services, and agentic tools, while timelines and safety risks are tightening as capability advances.
“AI will replace all the jobs!” Is Just Tech Execs Doing Marketing
The post argues that claims about AI replacing 20-50% of jobs are not supported by history or current labor data. It says AI is more likely to shift hiring and create jobs than to eliminate work at scale.
Agentic Design Systems in 2026
Brad Frost argues that design systems combined with AI can support collaborative product work, letting non-technical teammates shape features while AI stays constrained by production-grade components and standards.
The Impossible Backhand
The post argues that AI hits a structural quality ceiling, so deep domain expertise becomes more valuable, not less. Human-AI collaboration works best when people can spot where models are confidently wrong.
Does AI mean the demand on labor goes up?
The page argues that AI may increase, not reduce, labor demand. As tools become more capable, people do more work, raise expectations, and turn saved time back into work.
Balancing Code and Teams: How Microsoft Does Effective Engineering Management
Nick Cosentino describes engineering management as a people-first role built on situational leadership, transparency, trust, career alignment, safe delivery practices, and thoughtful use of AI.
Friction Was the Feature
The essay argues that AI makes words cheap, which floods hiring, admissions, reviews, and other systems with polished but low-value output. The fix is to replace prose with harder-to-fake proof and keep friction where it preserves signal.
Life as a physical process
The text proposes treating life as a physical process that generates systems statistically unlikely to assemble and persist, using Assembly Theory to rethink life, aliveness, and even AI.
LLMs are mirrors of operator skill
The author argues that AI is a matter of skill: LLMs reflect the operator’s level, which makes it harder to identify qualified operators and shows that the interview process is broken.
Knowledge Work Is Dying—Here’s What Comes Next
The piece argues that AI will commoditize knowledge work and raise the value of wisdom work built on emotional clarity, discernment, and connection, especially in leadership and teamwork.
Why Generalists Own the Future
The piece argues that AI weakens shallow specialist skill but strengthens generalists, who adapt across domains, ask the right questions, and use language models to work through novel problems.
Superintelligence is impossible
The essay argues that superintelligence is overblown because intelligence is not a single unlimited scale. Using the No Free Lunch principle, it says gains in one domain require losses in others.
Big Tech is replacing human artists with AI
The text argues that scaling laws will push advanced creative AI into Big Tech’s hands, making art, writing, music, and translation corporate-controlled and detached from consciousness.
The most valuable skill in 2026 isn't writing code. It is deleting it.
The text argues that AI makes creation cheap but maintenance costly, so senior engineering is shifting toward removing code, shrinking complexity, and curating systems instead of adding more.
Who Are We Still Writing Technical Articles For?
The author argues that technical articles still matter, but for a smaller, more attentive audience. He favors reflective, experience-based writing over tutorials, which age quickly and are often better handled elsewhere.
Why "AI is the new sustainable"
The article argues that AI is a vague, overused marketing term, warns about its energy use, unpaid labor, unreliability, and military misuse, and says it should be questioned rather than blindly embraced.
Why AI Can't Replace Developers: The Real Limits of AI in Coding
The text argues that AI helps with repetitive and simple tasks, but fails at context, reasoning and debugging complex problems; therefore, it should be treated as an assistant.
Do Frameworks Matter Anymore?
The post argues that frameworks still matter because they provide abstractions, structure, and constraints, and that new frameworks may still be worth building for an AI-assisted workflow.
A Stoic’s Guide to Navigating AI Hype Cycles
The text uses Stoicism to argue that durable AI products should focus on what is under control, accept the hype cycle, and prioritize utility, trust, and permanence.
Software engineering job openings hit five-year low?
Gergely Orosz uses Indeed data to show software developer job postings down 35% since January 2020, worse than any other tracked industry, and works through zero-interest-rate policy, Section 174, and AI as candidate causes.
How not to lose your job to AI
Benjamin Todd's 80,000 Hours framework for which work skills gain value as AI advances, built on why automation often raises wages before it cuts them.
Anthropic's co-founder wants an AI teddy bear for his kid
Anthropic co-founder Jack Clark says he'd buy an AI-powered teddy bear to entertain his kid during chores, as a supplement to human friendship.
News Feature: What are the limits of deep learning?
PNAS profile of 2019 surveys where deep learning breaks: adversarial spoofing, data inefficiency, opacity, and the missing common sense that symbolic AI once tried to hand-code.
A Primer on Large Language Models and their Limitations
Johnson and Hyland-Wood's LLM primer argues hallucination is the wrong word: LLMs produce Frankfurtian bullshit, indifferent to truth, alongside catastrophic forgetting, model collapse, and jailbreak risk.
A Functional Taxonomy of World Models - Dr. Fei-Fei Li
Fei-Fei Li and the World Labs team split 'world model' into three functions: renderers output pixels, simulators output physically faithful state, planners output actions.
Should Junior Developers Still Learn JavaScript the Hard Way?
Argues junior developers should still learn JavaScript fundamentals — closures, scope, the event loop — before frameworks, blaming boring teaching, not the material, for why it feels pointless.
The Slow and Quiet Cognitive Atrophy of a Modern Software Engineer
AI-generated code looks clean and its own tests pass, but skipping the review step quietly erodes an engineer's own skill. Elmar Chavez traces the decay and argues discipline is what keeps understanding intact.
How we built a desktop companion robot with Gemma 4 and Raspberry Pi
A LEGO-bodied desk robot switches between a local Gemma 4 model for instant private chat and cloud Gemini for harder reasoning, its eyes and motors driven by a five-state machine.
IC work is the new career flex
Elena Verna argues that AI makes high-impact individual contributor roles possible again: one person can now run projects end to end, reduce coordination overhead, and deliver business value without a team under them.
Write broken commits for better review
The article argues that intentionally broken commits can make code review easier when they separate mechanical changes from real ones, especially under squash merges and AI-assisted development.
Teaching Forces Clarity, AI Amplifies Clarity
00:22:11 — Teaching Forces Clarity, AI Amplifies Clarity
AI as Amplifier: Clarity or Dysfunction
00:10:07 — AI as Amplifier: DORA Report on Clarity vs Dysfunction
AI CEOs and public fear of AI
AI CEOs are warning about AI’s disruption in ways that may help fundraising and enterprise sales, but those same messages are making the public less trusting as elections near.