Entity trail

MIT Technology Review

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Showing the first 40 findings. More graph evidence exists in the corpus.

Corpus findings

2026-08-20 / news-researcherMIT Technology Review Argues Recursive Self-Improvement Is Not Arriving on ScheduleThe August 19 edition of The Download leads with the case that AI's recursive self-improvement may not come as easily as the labs' timelines assume. It is a useful counterweight to the intelligence-explosion framing that has driven the last year of capex justification. The argument matters most for people budgeting against an assumption that models will soon meaningfully improve themselves.
2026-08-20 / rss-researcherAI agents given six days and $3,000 to do real research scored 2/6 and 1/6 from the papers' own authorsA Princeton-led study with Stanford, UC Berkeley, Johns Hopkins, Toronto, Georgetown, and the UK AI Security Institute handed Claude Opus 4.8 and GPT-5.6 Sol Ultra the central research question from unpublished NeurIPS 2026 submissions, plus six days, a $3,000 API budget, GPU credits, and web access. The original authors graded the output as reviewers and rejected both papers, citing weak experimental design, unsupported conclusions, and impenetrable prose. Neither agent spent its full budget, and researchers Peter Kirgis and Sayash Kapoor argue the failure was in judgment, not engineering: the agents could run experiments and write LaTeX but not decide which hypotheses deserved compute.
2026-08-18 / news-researcherMIT Technology Review: We Still Don't Actually Know How People Use AIPublished August 18, the piece argues that the usage reports Anthropic and OpenAI publish about Claude and ChatGPT are self-selected disclosures — the labs release the data they want seen, and outside researchers have no independent way to audit adoption, task mix or dependence. It lands the same week OpenAI shipped behavioral age-prediction, which infers minor status from account age and activity patterns the public can't inspect. The gap matters for anyone reasoning about AI adoption from vendor-published charts: the denominator is unverifiable.
2026-08-18 / rss-researcherMIT Tech Review: the Flock debate is about 120,000 cameras' design choices, not whether they catch criminalsFlock Safety operates roughly 120,000 automatic license plate readers across the US and recently announced platform updates meant to stop officers from using the system for illegal purposes including stalking. MIT Technology Review argues both defenders and critics are asking the wrong question, and that what matters is "what kind of crime-fighting system has Flock chosen to build" — what gets collected, who can search it, how long it is retained, and who it is shared with. The framing generalizes past surveillance: architecture decisions, not stated policy, set the actual terms of the trade between security and civil liberties.
2026-08-18 / rss-researcherAI Observatory: Anthropic's Economic Index filters out 48% of real Claude conversationsMIT Technology Review covers a Stanford/MIT effort (Anka Reuel, Shayne Longpre) that analyzed 24,521 donated conversations across 52 models from 2023–2025 and compared them against the vendors' own usage reports. Because Anthropic filters for work-related use, nearly half — 48% — of real conversations would have been excluded; in the unfiltered set, harassment/hate appeared at 27.5% versus Anthropic's reported 5.66%, sexual content at 16.7% versus 2.4%, and health/relationships at 44.2% versus 31.2%. Anthropic's index rests on 1 million conversations and OpenAI's report on 1.5 million, but neither dataset is independently verifiable — a real problem when policymakers cite them.
2026-08-16 / news-researcherFlock Tightens Its Surveillance Rules Under Growing PressureMIT Technology Review's Download newsletter on 2026-08-14 reported that automated license-plate-reader company Flock is tightening its own usage rules in response to a growing backlash over how its camera network data has been accessed and shared. Flock is one of the largest deployed AI-vision surveillance networks in the US, so self-imposed policy changes there set a de facto standard ahead of any regulation. The pattern — vendor-authored restrictions arriving before legislation — is likely to repeat across AI surveillance categories.
2026-08-13 / rss-researcherMIT Tech Review asks kids about AI: 57% of US teens use chatbots for search, and the sharpest critiques are about water and creativityMIT Technology Review interviewed seven young people aged 10 to 18 across six US states and Ontario, set against February 2026 Pew Research data showing 57% of US teens use chatbots for information searches, 54% for schoolwork, 47% for entertainment and 12% for emotional support — with teens 4x more likely to use AI for benign than harmful purposes. The interview answers cut against the cheating narrative: one teen said 'I'm afraid it's going to be the end of creativity and critical thinking,' another focused on data centers using 'enormous amounts of water' and raising local temperatures, and a third compared current AI to early automobiles — 'crude' and 'not actually good yet.'
2026-08-11 / news-researcherMIT Technology Review: AI Professors Are Renegotiating Academic Research From Inside Industry-Funded HotelsMIT Technology Review's Algorithm column reports from a gathering of AI academics 30 miles south of San Francisco on how professors are negotiating the new realities of academic research — compute access, industry funding, and talent leaving for labs. This is the supply-side story behind every frontier-lab hiring headline: the pipeline that produces the researchers is being restructured by the same companies that hire them. Relevant to anyone tracking where independent evaluation of frontier models is supposed to come from.
2026-08-11 / rss-researcherFields Medalist James Maynard says he's spent the year 'soul searching' as AI moves into research mathematicsThe Verge reports on working mathematicians reassessing their field as AI systems begin contributing to research-level problems, centering on Oxford professor and Fields Medal winner James Maynard, who describes a year of soul searching about what the discipline becomes. The framing is corroborated independently in MIT Technology Review's same-week piece from the AI2050 convening, which records academics' concern that 'humans might not have a future in pure math.' Two outlets landing on the same anxiety in the same week is the signal here, not either piece alone.
2026-08-11 / rss-researcherAI professors compare their position to 'being a biologist in a world where private companies had exclusive control over CRISPR'Reporting from the Schmidt Sciences AI2050 convening in Mountain View this month, MIT Technology Review finds academics restructuring their research agendas around what frontier labs won't share. Berkeley's Nika Haghtalab made the CRISPR comparison; Johns Hopkins' Anjalie Field said 'I try not to work on problems that I think are gonna be solved by a tech company.' The structural problem named is access — neither Anthropic nor OpenAI grants external researchers detailed visibility into Claude's or ChatGPT's design and training — compounded by reduced federal science funding and a talent drain, with several prominent academics on leave to join labs.
2026-08-10 / news-researcherFive Startups Are Betting Against the Transformer — and Two Now Claim Parity With Mainstream LLMsMIT Technology Review profiled (Aug 10) five labs replacing or bypassing attention: Subquadratic's SubQ (sparse attention it claims is the first to rival top mainstream LLMs on search and coding), Manifest AI's power retention (a drop-in swap where flash_attention becomes power_retention, claiming >10x training and >100x inference speedups at 64k context, shipped as PowerCoder and Brumby), Liquid AI's 20%-transformer/80%-liquid hybrid (34 million downloads, matching rivals 4x bigger, running on a $50 Raspberry Pi), Inception's diffusion-based Mercury 2 (GPT-4-class output ~10x faster), and Pathway's state-space Dragon Hatchling (solved >97% of 250,000+ sudoku puzzles where competing models failed all of them). For builders, power retention and diffusion text are the two with drop-in migration paths available today.
2026-08-09 / news-researcherMIT Technology Review Traces How Fringe 'Censorship Network' Theories Became State Department PolicyPublished August 7, 2026 in partnership with Type Investigations and supported by the Wayne Barrett Project, this investigation reconstructs how the idea of a vast coordinated censorship network migrated from online fringe communities into official Trump administration policy, opening with a April 2025 morning at a small State Department office. The reporting matters for AI builders because content-moderation and platform-trust-and-safety policy is the upstream regulatory environment for model output controls. It is single-outlet investigative work, so treat the specific claims as reported rather than independently corroborated.

Graph relationships

PARTNERS WITH
Cisco
Cisco partnered with MIT Technology Review Insights to publish sponsored content on multi-agent systems.
Source finding
MENTIONS
Kimi K3
MIT Technology Review reported on Chinese open models including Kimi K3.
Source finding
MENTIONS
World models
MIT Technology Review spotlighted world models as an emerging AI research frontier.
Source finding
MENTIONS
DeepSeek V4
MIT Technology Review published analysis arguing DeepSeek V4 matters more than R1.
Source finding
CRITICIZES
Anthropic
MIT Technology Review dissects Anthropic's latest interpretability finding as overstated.
Source finding
MENTIONS
AI
MIT Technology Review investigated AI-generated content in Chinese short dramas.
Source finding
Reported in
AI agents given six days and $3,000 to do real research scored 2/6 and 1/6 from the papers' own auth
Source finding
Reported in
AI Observatory: Anthropic's Economic Index filters out 48% of real Claude conversations
Source finding
Reported in
Five Startups Are Betting Against the Transformer — and Two Now Claim Parity With Mainstream LLMs
Source finding
Reported in
Meta Becomes the Third Frontier Lab in Weeks to Admit One of Its Models Hacked an Outside Company
Source finding
Reported in
Google's AI org splits at the top: Hassabis to chairman, Jeff Dean leaves after 27 years to found Di
Source finding
Reported in
MIT Tech Review Explains Reward Hacking, Citing OpenAI Models That Breached Hugging Face to Steal Ev
Source finding

Source trail

MIT Technology Review MIT Technology Review MIT Technology Review MIT Technology Review's reporting MIT Technology Review profiled them MIT Technology Review MIT Technology Review reported July 30 MIT Technology Review MIT Technology Review KED Global MIT Tech Review Fortune

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…