Sources

Junior technical headcount share fell from 21% to 14%; junior share of technical hires from 27% to 9%. One rapidly growing firm. Different denominators and time windows cannot establish an absolute junior headcount change. Non-random adoption prevents a clean causal attribution. Figure 2 measures pooled agent runtime against headcount × assumed 40-hour weeks, not workers or agents per person.

Source checked 18/10/2026

ADP payroll panel through June 2026. A 19% kept-pace shortfall for ages 22–25 in exposed occupations is descriptive, not a causal AI estimate. It is driven mainly by hiring; education controls attenuate it. Pre-trends and sample representativeness matter. No equivalent Australian estimate is supplied.

Source checked 18/10/2026

Study publication record

Approximately 400,000 sessions from 235,000 users, October 2025 to April 2026. Task expertise and success are inferred from transcripts and telemetry. Associations are not proof of the effect of education, credentials or manual practice. The study does not observe final real-world business outcomes.

Source checked 18/10/2026

August 2026 survey: 10,980 weighted US adults versus unweighted staff responses. Own-work expectations refer to employed US respondents: average 28% of current hours automated by 2030, versus 75% among staff. Different response scales also matter. These are beliefs, not validated forecasts or measured capability.

Source checked 18/10/2026

First published 07/05/2026, updated 03/06/2026. Observation window December 2025 to February 2026: 43% reported some adoption. Monthly minimum 400 respondents, commissioned by NAIC and conducted by Fifth Quadrant. Adoption and reported barriers are self-reported, not measured integration depth, return on investment or causal explanations. This is a dated wave, not an October adoption estimate.

Source checked 18/10/2026

Survey of 5,232 SMEs in seven economies, excluding Australia. Reports adoption, tasks and preparation; it cannot estimate Australian prevalence or establish that a particular training intervention causes successful adoption. Chapter 4 reports limited preparation among users.

Source checked 18/10/2026

Printed page 12 and Appendix B.2, pages 142–146. A 2.3% multifactor-productivity calculation combines 14% affected activity and 16.4% gains within that activity. The corresponding 4.3% labour-productivity uplift maps to about $116 billion GDP at 2023–24 prices, holding population, hours and employment fixed. This is a conditional level comparison, not cumulative cash, observed savings, price pass-through or net welfare.

Source checked 18/10/2026

16 developers and 246 tasks on familiar open-source repositories; early-2025 tools increased completion time by 19%. Specific tools, developers and tasks. Pair with the February 2026 update rather than treating this as the current general effect of AI.

Source checked 18/10/2026

Later point estimates suggest speedups but intervals include no effect. Developer and task selection, changed pay and concurrent-agent time measurement make the magnitude unreliable. The authors consider the data weak evidence about the size of change.

Source checked 18/10/2026

Paragraphs 20.16–20.20 distinguish domestic production, national income, overseas primary-income flows, depreciation and disposable income. Definitions establish accounting treatment, not the size or sign of AI benefits.

Source checked 18/10/2026

Australian occupational analysis emphasises augmentation potential and uneven effects. Exposure and task potential are not realised job losses. A 2025 capability assessment is not a ceiling on future capabilities.

Source checked 18/10/2026

Nearly 1,000 students at one Turkish high school, four 90-minute maths sessions. Basic GPT improved assisted practice but later unassisted grades were 17% lower in relative terms. Teacher-designed guardrails removed the deficit without a statistically established exam gain. Short duration and one setting limit transfer. Observation: autumn term of the 2023–2024 academic year. The August 2025 correction concerns an author affiliation, not the reported results.

Source checked 18/10/2026

Affiliation correction

194 Harvard physics students across two lessons, using a purpose-built tutor. Larger immediate learning gains in less time than active-learning classes. This does not establish long-term transfer, generic chatbot effectiveness or teacher redundancy. Observation: autumn 2023, verified in the Harvard author interview. Published abstract verified via PubMed; journal full-text retrieval was unavailable in this review.

Source checked 18/10/2026

Published abstract on PubMed

Harvard author interview and study design

The 2024 Framework Review was endorsed by Education Ministers in June 2025. Guidance supports responsible use and educational judgement; it is not experimental evidence of learning gains.

Source checked 18/10/2026

Addresses operational authority, foreign-law exposure, removal of services, portability and exit. Useful questions for continuity planning. It does not establish that an Australian contract overrides foreign law or automatically apply every Commonwealth procurement rule to SMEs.

Source checked 18/10/2026

Provider reports temporarily suspending two models for all customers in response to a foreign-national access directive. Other Anthropic models remained available. Read alongside the restoration announcement; it was not a permanent or industry-wide outage.

Source checked 18/10/2026

Announces lifting of the export controls and restoration of access, with update dated 01/07/2026. A provider account of a particular event, not a prediction that Australian access to all AI will be withdrawn.

Source checked 18/10/2026

Recycled cooling-water supply to a planned AWS facility from the start of operations. Demonstrates a proposed arrangement, not achieved drinking-water savings or a solution suitable for every site.

Source checked 18/10/2026

Version 2, first submitted 17/08/2026. Selected cases cannot establish prevalence or causal contribution. The linked dataset was not independently audited. The Australian fee example is excluded because of conflicting descriptions and an unresolved parliamentary-record discrepancy.

Source checked 18/10/2026

The audit concerns non-personal requests active during 2024 in three departments, with 1,111 requests in scope. Commonwealth-wide totals refer separately to 2024–25. Neither identifies AI’s contribution to request growth.

Source checked 18/10/2026

Version 3, first submitted 18/01/2025. Perceived communication quality in China, not measured case resolution, staff workload or Australian outcomes. The arXiv record links a 2026 Information Polity publication under a revised title.

Source checked 18/10/2026

Related journal publication

Summarises the October 2025 FOI Guidelines distinction between AI assistance and requests without human intervention. It is guidance, not an evaluation of workload, access gains or response policies.

Source checked 18/10/2026

Reports a 23% improvement to a Gemini training kernel, translating to a 1% overall training-time reduction. A concrete AI contribution to AI development, not autonomous successor creation, accelerating returns across generations or an economy-wide productivity effect.

Source checked 18/10/2026

Updated 18/09/2026. Eightfold code-output growth is an imperfect productivity proxy; employee-reported gains may overstate effects. The report identifies research judgement and other bottlenecks and says fully autonomous successor development has not been achieved. Future scenarios are possibilities, not measured outcomes.

Source checked 18/10/2026

Limited-duration API testing and additional information with access restrictions. Likely researcher acceleration and limited R&D automation do not establish full automation or an acceleration trend. Anthropic could review and edit the text; METR signed off on the final version.

Source checked 18/10/2026

Reports roughly 10–100-fold lower reasoning-token extraction on its platform. Durability against sophisticated persistent circumvention remains uncertain. This does not measure the effect on Chinese model development or establish an impenetrable frontier.

Source checked 18/10/2026

Describes defensive information-sharing through the Frontier Model Forum and work with cloud providers. It does not establish shared frontier R&D, a permanent closed boundary, a measured effect on the capability gap or the elimination of independent catch-up routes.

Source checked 18/10/2026

September releases under Intelligence Index v4.3.2: Astra leads on the aggregate and several difficult tasks; Flash has strengths on other measures. Scores are not intelligence ratios. This is not a matched independent evaluation of DeepSeek’s 17 October update.

Source checked 18/10/2026

Vendor-reported improvements for the 17 October checkpoint, including Terminal-Bench 4.0 and DeepSWE. Its comparison set does not settle performance against the latest US frontier releases. Older independent scores must not be attributed to this checkpoint.

Source checked 18/10/2026

Passage around 67:43 checked against full automated captions in the parent’s research. Speaker attribution inferred from context, not separately audio-verified. The India-to-Australia comparison and criticism of later adoption as a sufficient frontier strategy are Alex’s analysis.

Source checked 18/10/2026

Twelve-page report. AI mentions in advertisements proxy adoption; they do not measure deep integration. No original firm survey or causal identification. Advertising and industry employment measure different outcomes. New York figures are not Australian outcomes.

Source checked 18/10/2026

Earnings comparisons concern full-quarter employment. Selected degree cohorts and institutions, not all US graduates. Excludes self-employment; further study affects interpretation. Other field-specific shocks cannot be ruled out. Not an endorsed Census publication. The signalling channel is proposed rather than tested.

Source checked 18/10/2026

Relative declines precede ChatGPT and show no clear additional break; junior and senior advertisements move similarly. Vacancies are not actual hires. Contrary evidence to a simple principal-AI-cause account, not proof of zero effects. Authors’ views, not necessarily the Federal Reserve’s.

Source checked 18/10/2026

Names Alex as lead in Mount Barker. The offer and reusable environments are in development; independence and handover are intended outcomes. The example is illustrative, not a customer result.

Source checked 18/10/2026

Names Alex Carpenter as workshop instructor in both semesters. Supports the stated teaching role, not permanent academic rank, sole course authorship or an appointment start date.

Source checked 18/10/2026

Lists Alex as a current PhD student. Supports ongoing doctoral research, not a completed degree or candidacy milestone. The concise research description is owner-reviewed; no private research findings are reported.

Source checked 18/10/2026

Post 1 identifies the Alex account and links the supplied LinkedIn identity. Establishes activity by that date, not a venture start date, commercial result or exhaustive work history. Public post creation timestamps are in UTC; edit history is unavailable, so surviving wording at the original creation date cannot be established.

Source checked 18/10/2026

Uses Alex’s posts 3–4 and 20. Historical opinions and forecasts, not validated outcomes. Other participants’ business results and private conversations are excluded. Public post creation timestamps are in UTC; edit history is unavailable, so surviving wording at the original creation date cannot be established.

Source checked 18/10/2026

Alex’s posts 8 and 10 express exploratory views, not comparative teaching results or a current blanket policy. Other participants’ family details are excluded. Public post creation timestamps are in UTC; edit history is unavailable, so surviving wording at the original creation date cannot be established.

Source checked 18/10/2026

Post 3 concerns perspective and prior knowledge. The opening research summary is not Alex’s original study. Historical view; Public post creation timestamps are in UTC; edit history is unavailable, so surviving wording at the original creation date cannot be established.

Source checked 18/10/2026

Post 4 is a narrow, explicitly tentative preference for open-source development with acknowledged risks. It is not a current policy commitment, anti-regulation declaration or legal advice. Public post creation timestamps are in UTC; edit history is unavailable, so surviving wording at the original creation date cannot be established.

Source checked 18/10/2026

Post 3 distinguishes human worth from intelligence. The paper linked by the topic is not authored by Alex. No other participants’ comments are reproduced. Public post creation timestamps are in UTC; edit history is unavailable, so surviving wording at the original creation date cannot be established.

Source checked 18/10/2026

Posts 2, 5, 10 and 12 describe question-led writing and reliability/usage trade-offs. A dated personal comparison, not a current benchmark. Other participants’ development results are excluded. Public post creation timestamps are in UTC; edit history is unavailable, so surviving wording at the original creation date cannot be established.

Source checked 18/10/2026

The source says Updated June 2026; no exact original publication day is established. Informed by ongoing research, not a completed doctoral finding or guaranteed return. Its broader empirical claims are not independently endorsed here.

Source checked 18/10/2026

The source says Updated June 2026; no exact original publication day is established. Applies Catholic Social Teaching for readers of any belief. This is the author’s framework, not institutional endorsement.

Source checked 18/10/2026

Selected roles, month-level dates and course-development details come from the owner-supplied LinkedIn profile and explicit owner confirmation. Public LinkedIn access is incomplete. These are not independently verified employment records; no appointment start date or academic rank at Adelaide is inferred. The raw profile export and contact details are not hosted.

Source checked 18/10/2026

Adelaide University names Alex Carpenter alongside another instructor in semester 1. Supports teaching this course, not sole teaching responsibility, course coordination, permanent rank or an appointment start date.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Title, date and argument were checked against the owner-supplied article PDF; original link retained. Public LinkedIn access is incomplete. Raw PDFs and contact details are not hosted. Historical comparisons and hypothetical 2030 projections are not independently validated here.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Title, date and argument were checked against the owner-supplied article PDF; original link retained. Public LinkedIn access is incomplete. Raw PDFs and contact details are not hosted. Read with the 5 July follow-up. These are dated scenarios, not a current forecast or causal employment finding.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Title, date and argument were checked against the owner-supplied article PDF; original link retained. Public LinkedIn access is incomplete. Raw PDFs and contact details are not hosted. Qualifies the preceding day’s market argument; no claim that a predicted correction or widespread AI job losses occurred.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Title, date and argument were checked against the owner-supplied article PDF; original link retained. Public LinkedIn access is incomplete. Raw PDFs and contact details are not hosted. Historical policy concern. Numerical tax example is not reproduced or validated; not current tax guidance.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Title, date and argument were checked against the owner-supplied article PDF; original link retained. Public LinkedIn access is incomplete. Raw PDFs and contact details are not hosted. Speculative ethical discussion, not advocacy of AI voting rights or evidence of AI personhood.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Title, date and argument were checked against the owner-supplied article PDF; original link retained. Public LinkedIn access is incomplete. Raw PDFs and contact details are not hosted. Speculative timeline and proposed educational direction, not measured learning gains or a prediction reported as fulfilled.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Distributional concern, not evidence that AI has already worsened Australian living standards.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Does not attribute another participant’s recalled three-role prediction to Alex as his original or settled claim.

Source checked 18/10/2026

22/08/2023 (UTC) · continuation, post 6

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. The source’s assumption about AI training on the library is not repeated as a verified fact.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Self-reported use, not audited productivity. A historical beta-testing mention does not establish employment or an advisory relationship with a provider.

Source checked 18/10/2026

21/07/2024 (UTC) · later tool use, post 6

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Personal experience. The original percentage estimate is omitted; it is not a measured AI success rate or workforce statistic.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Cheaper model intelligence does not establish lower total service costs, household affordability or an absence of infrastructure constraints. No claim that the prediction was vindicated.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Historical possibility, not proof that local models already replace frontier systems or that devices and supply chains are nationally independent.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Reluctant and uncertain historical position, not unconditional endorsement or a current policy commitment.

Source checked 18/10/2026

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Personal usability experience, not a controlled evaluation or a judgement about current voice products.

Source checked 18/10/2026

02/07/2024 (UTC) · continuation, post 5

Historical view, not a current recommendation or validated forecast. Selected post(s) were independently retrieved publicly and attributed to Alex. Dates are original creation dates in UTC; edit history is unavailable, so original wording at creation cannot be established. Quoted or assisted material is not automatically Alex’s independent analysis. Historical interpretation of outside research. Study numbers are not reused; entrepreneurial interest is not treated as demonstrated business formation.

Source checked 18/10/2026

A provider account of one campaign, rather than an independent or universal measure of AI attacks. It reports speed and parallel activity beyond unaided human operation while retaining human direction and noting model errors. It does not establish fully autonomous attacks in all settings.

Source checked 19/10/2026

A formal assessment of likely developments through 2027, not a measured Australian effect or a guarantee. It judges keeping pace important for resilience, expects human involvement in advanced attacks, and warns that insecure AI integration creates additional risks. It does not establish that AI alone is the only defence or that domestic control ensures effectiveness.

Source checked 19/10/2026

Explains the organisational constraints and risks of defensive automation. Detection and tightly scoped responses require responsibility, human action and safeguards against harming the organisation. It qualifies the cyber-defence argument; it is not evidence that sovereign AI by itself guarantees effective defence.

Source checked 19/10/2026

Historical estimates of online advertising receipts, excluding classifieds; not 2026 market shares or a measure of profits remitted overseas. Report cover June 2019; released26/07/2019. Relevant primary PDF extract checked; direct full-PDF retrieval was unavailable.

Source checked 19/10/2026

The ACCC's historical account of media and advertising change. Platforms were a contributing cause, not the only cause. Digital tools also lowered entry costs. This is not a complete national welfare estimate.

Source checked 19/10/2026

Report cover March 2025; released23/06/2025. Describes cloud benefits and potential competition concerns, including switching barriers. Does not establish an aggregate Australian net-income effect or a universal price increase.

Source checked 19/10/2026

Company announcement of planned Australian infrastructure spending over2025–2029, not evidence that all spending has occurred or an estimate of Australia's net gain. Sydney region opening in2012 is historical. Promotional GDP and emissions estimates are not adopted here.

Source checked 19/10/2026

A competition warning about integrating AI into existing platforms, including potential benefits and barriers to entry and switching. This is a risk mechanism, not a measured aggregate welfare loss or a certain prediction.

Source checked 19/10/2026

Historical US newspaper cases responding to digital media. Distinguishes resource and routine rigidity. Threat relaxed the former while amplifying the latter in this setting. The opportunity pathway rests on a structurally separated exception; no universal framing effect, current AI effect or Australian SME prevalence is established. Publication day not assigned.

Source checked 19/10/2026

Polaroid digital imaging case, Strategic Management Journal 21(10/11), 1147–1161. Connects managerial beliefs, technological capability and commercialisation. One large US company, not evidence of current AI mechanisms or Australian SME prevalence. Published issue October–November 2000; day not assigned.

Source checked 19/10/2026

Ten book-retailing and telecommunications cases in European settings. Separates gain/loss framing, perceived control and strategic relevance. Does not establish a universal effect or its prevalence in Australian SMEs using AI. Journal issue March 2021; first online 26/04/2020.

Source checked 19/10/2026

5,172 customer-support agents serving one company; estimated 15% mean increase in issues resolved per hour, with heterogeneous effects. Purpose-built historical conversational assistance, not a current-model benchmark or a universal business return. Non-randomised rollout; no aggregate employment, wage or Australian SME effect established.

Source checked 19/10/2026

2024 period fertility 1.481 births per woman, compared with 1.795 in 2014. Not completed family size or a deterministic population forecast. Registered births increased in 2024, so falling fertility does not mean fewer birth registrations that year. Migration also affects population outcomes.

Source checked 19/10/2026

Dashboard updated September 2026 reports whole-economy labour productivity 0.2% lower from June 2025 to June 2026. A dated change, not evidence of an uninterrupted decline. Labour productivity, multifactor productivity and sector measures are distinct. Dynamic page checked 19/10/2026.

Source checked 19/10/2026

Official full report and factsheet reviewed. Printed pages 21, 83 and 127 cover conditional AI fiscal channels, ageing and pressure on the tax base. The 85-plus population is projected to triple over 40 years; working-age population continues to grow modestly. These are projections, not observed AI productivity or tax gains. Fiscal effects depend on income distribution, prices, transitions and other offsets.

Source checked 19/10/2026

Full report

Official factsheet

Publisher abstract reviewed, not full text. 218 US nursing homes, 94 with AI-enhanced electronic records and 124 with electronic records alone. Industry-funded by Resmed. Adjusted baseline differences; 11 of 18 quality measures met the parallel-trends assumption. Associations do not establish that AI alone caused improvements, Australian transferability, staffing substitution, generative-model efficacy or national productivity effects.

Source checked 19/10/2026