Alex Carpenter · AI in practice

Alex Carpenter · AI in practiceNew and revised editions of living articles on AI, work and Australia.https://www.alexcarpenter.au/2026-10-20T12:00:00ZAlex CarpenterAbout Alex · v5https://www.alexcarpenter.au/about/versions/2026-10-20-v5/2026-10-20T12:00:00Z

Identify the encyclical the site’s values draw on, at the author’s request.

About Alex · v4https://www.alexcarpenter.au/about/versions/2026-10-20-v4/2026-10-20T12:00:00Z

Shorten the page and align its presentation with the revised articles, using the wording reviewed and approved by Alex on 20/10/2026. Remove Entrepreneurial Business Development from the current About page; retain the approved personal perspective and career history.

Public record · v3https://www.alexcarpenter.au/public-record/versions/2026-10-20-v3/2026-10-20T12:00:00Z

Shorten the page and align its presentation with the revised articles, using the wording reviewed and approved by Alex on 20/10/2026. Preserve historical positions, dates, originals and the July 2024 reconsideration; update current-article links.

Will Australians gain once we count the costs and overseas income? · v6https://www.alexcarpenter.au/articles/australian-net-benefits/versions/2026-10-19-v6/2026-10-19T12:00:00Z

Simplified the summary and closing after reviewing the nine articles together. Retains my uncertainty about how Australia should secure and share the overall gains.

What happens to the junior career ladder when AI does the first draft? · v6https://www.alexcarpenter.au/articles/junior-career-ladder/versions/2026-10-19-v6/2026-10-19T12:00:00Z

Removed a repeated introduction and focused the shared customer-support study on novice performance and learning, rather than repeating its full numerical result.

What should students learn when AI can help with almost everything? · v6https://www.alexcarpenter.au/articles/what-students-should-learn/versions/2026-10-19-v6/2026-10-19T12:00:00Z

Removed the final paragraph because it exactly repeated the page summary. The argument and evidence are unchanged.

Will Australians gain once we count the costs and overseas income? · v5https://www.alexcarpenter.au/articles/australian-net-benefits/versions/2026-10-19-v5/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

Why can AI do so much while business change takes so long? · v5https://www.alexcarpenter.au/articles/capability-and-adoption/versions/2026-10-19-v5/2026-10-19T12:00:00Z

Removed repeated framing around the historical research and the purpose of my own research. The historical studies, AI evidence and substantive limits are unchanged.

When AI makes requests cheaper, who pays to answer them? · v5https://www.alexcarpenter.au/articles/cheaper-requests/versions/2026-10-19-v5/2026-10-19T12:00:00Z

Removed the evaluation principle repeated from the page summary and moved directly from the evidence to the proposed trials. The evidence and closing argument are unchanged.

What happens to the junior career ladder when AI does the first draft? · v5https://www.alexcarpenter.au/articles/junior-career-ladder/versions/2026-10-19-v5/2026-10-19T12:00:00Z

Removed a redundant caveat from the article summary at the author’s request. The answer’s substance, evidence and confidence are unchanged.

What should students learn when AI can help with almost everything? · v5https://www.alexcarpenter.au/articles/what-students-should-learn/versions/2026-10-19-v5/2026-10-19T12:00:00Z

Rewrote the argument around subject knowledge, learning design and assessment. Connects classroom experiments and task-expertise evidence with practical ways to test what students understand.

What would Australian control of AI actually mean? · v4https://www.alexcarpenter.au/articles/australian-control/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Simplified the structure and removed repetition. Added my assessment of Australia’s unproven continuity for frontier models and successive generations. Sources, historical editions and the article’s confidence assessment are unchanged.

Can Australia benefit from AI without controlling the frontier? · v4https://www.alexcarpenter.au/articles/australian-frontier-dependence/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Clarified the distinction between income from resources and data centres and Australia’s overall outcome. Replaced the repeated access-interruption chronology with a link to the control article, clarified what my 99% forecast refers to, and removed a closing paragraph already stated in the summary.

Will Australians gain once we count the costs and overseas income? · v4https://www.alexcarpenter.au/articles/australian-net-benefits/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Simplified the argument, made my uncertainty explicit, and added dated evidence on digital advertising, journalism, cloud services, Australian investment and AI platform dependence. These examples frame the problem without asserting a complete net benefit or a settled policy solution.

Why can AI do so much while business change takes so long? · v4https://www.alexcarpenter.au/articles/capability-and-adoption/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

When AI makes requests cheaper, who pays to answer them? · v4https://www.alexcarpenter.au/articles/cheaper-requests/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Rewrote the article around evidence of better outcomes, unnecessary work and AI-assisted processing. Adds a proposal for whole-service trials and the practical need for AI assistance as AI-generated workloads grow.

When should the public help pay for AI infrastructure? · v4https://www.alexcarpenter.au/articles/infrastructure-and-public-benefit/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Removed the customer-support result already covered in the adoption article and sharpened the test for public support of an Australian facility.

What happens to the junior career ladder when AI does the first draft? · v4https://www.alexcarpenter.au/articles/junior-career-ladder/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

If AI raises productivity, will everyday life become more affordable? · v4https://www.alexcarpenter.au/articles/productivity-and-affordability/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Removed the final paragraph because it exactly repeated the page summary. The argument and evidence are unchanged.

What should students learn when AI can help with almost everything? · v4https://www.alexcarpenter.au/articles/what-students-should-learn/versions/2026-10-19-v4/2026-10-19T12:00:00Z

Removed a redundant caveat from the article summary at the author’s request. The answer’s substance, evidence and confidence are unchanged.

What would Australian control of AI actually mean? · v3https://www.alexcarpenter.au/articles/australian-control/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Reframed the opening as an exploration, made continuity one test, and added the owner-approved cyber-defence and evolving-capability argument. Removed current-argument appeals to historical submissions and the two closing sections at the owner’s request. Historical records remain unchanged.

Can Australia benefit from AI without controlling the frontier? · v3https://www.alexcarpenter.au/articles/australian-frontier-dependence/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Rewrote the article around Australia’s expected resource and data-centre gains, pursuit of control over multiple frontier models, and acceptance of the best available arrangement. Adds the author’s self-improvement forecast and distillation assessment, with a documented local-inference example.

Why can AI do so much while business change takes so long? · v3https://www.alexcarpenter.au/articles/capability-and-adoption/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Rewrote the argument around resource commitments, operating routines and managerial interpretations, supported by historical primary studies and direct AI evidence. Distinguishes illustrative applications from findings about Australian SMEs and includes no Paper 3 findings.

When AI makes requests cheaper, who pays to answer them? · v3https://www.alexcarpenter.au/articles/cheaper-requests/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Removed a redundant caveat from the article summary at the author’s request. The answer’s substance, evidence and confidence are unchanged.

When should the public help pay for AI infrastructure? · v3https://www.alexcarpenter.au/articles/infrastructure-and-public-benefit/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

What happens to the junior career ladder when AI does the first draft? · v3https://www.alexcarpenter.au/articles/junior-career-ladder/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Simplified the argument around entry work, learning and progression. Balances descriptive US hiring warnings and contrary vacancy evidence with direct evidence that AI can help beginners. Australian effects remain uncertain.

If AI raises productivity, will everyday life become more affordable? · v3https://www.alexcarpenter.au/articles/productivity-and-affordability/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Reframed the question around affordability if AI delivers substantial productivity gains. Connects weak links, physical inputs, ownership, wages and Australia’s resource endowment, and adds my view that the tax code should secure a greater return for citizens.

What should students learn when AI can help with almost everything? · v3https://www.alexcarpenter.au/articles/what-students-should-learn/versions/2026-10-19-v3/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

Can Australia benefit from AI without controlling the frontier? · v2https://www.alexcarpenter.au/articles/australian-frontier-dependence/versions/2026-10-19-v2/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

When AI makes requests cheaper, who pays to answer them? · v2https://www.alexcarpenter.au/articles/cheaper-requests/versions/2026-10-19-v2/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

When should the public help pay for AI infrastructure? · v2https://www.alexcarpenter.au/articles/infrastructure-and-public-benefit/versions/2026-10-19-v2/2026-10-19T12:00:00Z

Rewrote the argument to connect AI infrastructure with ageing, low fertility and weak productivity. Adds official demographic and fiscal evidence and qualified early AI results, while retaining project-specific public-benefit tests and transparent contributions from governments and citizens.

Will AI productivity make everyday life more affordable? · v2https://www.alexcarpenter.au/articles/productivity-and-affordability/versions/2026-10-19-v2/2026-10-19T12:00:00Z

Removed redundant caveats and editorial self-reference at the author’s request. Evidence, positions and confidence are unchanged.

About Alex · v3https://www.alexcarpenter.au/about/versions/2026-10-18-v3/2026-10-18T12:00:00Z

Owner approved the exact first-person paragraph on 18/10/2026. It states his Catholic faith and family perspective. Earlier local drafts were not published on this site.

Will Australians gain once we count the costs and overseas income? · v3https://www.alexcarpenter.au/articles/australian-net-benefits/versions/2026-10-18-v3/2026-10-18T12:00:00Z

Added frontier dependence as a conditional influence on bargaining power and income distribution, with a link to Alex's new personal assessment. The overall sign of Australia's net benefit remains unresolved.

What would Australian control of AI actually mean? · v2https://www.alexcarpenter.au/articles/australian-control/versions/2026-10-18-v2/2026-10-18T12:00:00Z

Added a distinction between continuity for an agreed model and sustained access to an advancing frontier, with related analysis of AI-assisted research. The established objective of Australian-controlled inference remains unchanged.

Will Australians gain once we count the costs and overseas income? · v2https://www.alexcarpenter.au/articles/australian-net-benefits/versions/2026-10-18-v2/2026-10-18T12:00:00Z

Requested related reading added on 18/10/2026 alongside the new article. This revision adds a connection without changing the existing answer or source assessment.

Why can AI do so much while business change takes so long? · v2https://www.alexcarpenter.au/articles/capability-and-adoption/versions/2026-10-18-v2/2026-10-18T12:00:00Z

Added portability and provider dependence to the practical adoption discussion. Preserves the distinction between proposed evaluation criteria and empirical findings from the ongoing research.

What happens to the junior career ladder when AI does the first draft? · v2https://www.alexcarpenter.au/articles/junior-career-ladder/versions/2026-10-18-v2/2026-10-18T12:00:00Z

Owner-approved update on 18/10/2026 following the New York report assessment. Adds PFNYC, Census and New York Fed evidence and rechecks the latest Stanford study. Separates entry difficulties from proven AI causation.

What should students learn when AI can help with almost everything? · v2https://www.alexcarpenter.au/articles/what-students-should-learn/versions/2026-10-18-v2/2026-10-18T12:00:00Z

Requested related reading added on 18/10/2026 alongside the new article. This revision adds a connection without changing the existing answer or source assessment.

Public record · v2https://www.alexcarpenter.au/public-record/versions/2026-10-18-v2/2026-10-18T12:00:00Z

Owner-requested addition on 18/10/2026, incorporating the approved professional history, university work and selected public contributions. Earlier local drafts were not published on this site.

What would Australian control of AI actually mean? · v1https://www.alexcarpenter.au/articles/australian-control/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Launch of the living-article series, grounded in the dated submissions and primary sources listed here.

Can Australia benefit from AI without controlling the frontier? · v1https://www.alexcarpenter.au/articles/australian-frontier-dependence/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Requested on 18/10/2026. Sets out the concern that AI-assisted research may compound frontier advantage, distinguishes that outlook from demonstrated results, and separates useful adoption from ownership, operational continuity and assured frontier access.

Will Australians gain once we count the costs and overseas income? · v1https://www.alexcarpenter.au/articles/australian-net-benefits/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Launch of the living-article series, grounded in the dated submissions and primary sources listed here.

Why can AI do so much while business change takes so long? · v1https://www.alexcarpenter.au/articles/capability-and-adoption/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Launch of the living-article series, grounded in the dated submissions and primary sources listed here.

When AI makes requests cheaper, who pays to answer them? · v1https://www.alexcarpenter.au/articles/cheaper-requests/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Requested on 18/10/2026 to answer who pays for cheaper AI-assisted requests and how much additional demand may be a benefit previously missed. Uses the flooding paper, ANAO audit, communication study and OAIC guidance with their limits.

When should the public help pay for AI infrastructure? · v1https://www.alexcarpenter.au/articles/infrastructure-and-public-benefit/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Launch of the living-article series, grounded in the dated submissions and primary sources listed here.

What happens to the junior career ladder when AI does the first draft? · v1https://www.alexcarpenter.au/articles/junior-career-ladder/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Launch of the living-article series, grounded in the dated submissions and primary sources listed here.

Will AI productivity make everyday life more affordable? · v1https://www.alexcarpenter.au/articles/productivity-and-affordability/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Launch of the living-article series, grounded in the dated submissions and primary sources listed here.

What should students learn when AI can help with almost everything? · v1https://www.alexcarpenter.au/articles/what-students-should-learn/versions/2026-10-18-v1/2026-10-18T12:00:00Z

Launch of the living-article series, grounded in the dated submissions and primary sources listed here.