AI Tool Reviews in 2026: What Users Say About 8 Popular Tools
Public signals on Claude, ChatGPT, Cursor, GitHub Copilot, Perplexity, Consensus, Midjourney and Runway
By TheBestSuperintelligence Editorial TeamUpdated
What do users actually say about Claude, ChatGPT, Cursor and other AI tools in 2026?
Public discussion of eight AI tools, read in October 2026, centres on cost changes, source accuracy and privacy more than raw capability. Billing drew complaints for Cursor, ChatGPT and GitHub Copilot; Perplexity and Consensus draw citation warnings. These are signals, not test results.
☰On this page (8 sections)
- 01.Which 8 AI tools does this cover?
- 02.What do the signals say overall?
- 03.AI tool pricing complaints: Cursor, ChatGPT and GitHub Copilot
- 04.AI search citation accuracy: Perplexity and Consensus
- 05.AI tool privacy concerns: Claude, ChatGPT and Cursor
- 06.Tool by tool: what users praise, complain about and disagree on
- 07.Limits of this guide
- 08.Sources & Further Reading
Which 8 AI tools does this cover?
- 1ClaudePraised for Sonnet 5.5 price-performance; criticised for outages, safety filters and cost of heavy useChat assistantWindow: 2026-09-28 to 2026-10-18
- 2ChatGPTThe Pro plan change, ads and the Sites feature drew the most commentChat assistantWindow: 2026-08-30 to 2026-10-01
- 3CursorBilling visibility, ownership after the SpaceX deal and the agent-first Cursor 3AI code editorWindow: 2026-04-02 to 2026-10-18
- 4GitHub CopilotUsage-based billing from 1 June 2026 dominated the discussionCoding assistantWindow: 2026-04-20 to 2026-10-18
- 5PerplexityCitations that do not always support the sentence they are attached toAI searchWindow: 2025-03-06 to 2026-10-18
- 6ConsensusUseful for scoping; library guides warn against treating its Meter as evidenceResearch searchWindow: 2024-03-20 to 2026-10-18
- 7MidjourneyLittle first-hand image feedback; discussion centred on its medical-imaging moveImage generationWindow: 2026-05-11 to 2026-10-08
- 8RunwayFew user reviews; the documented criticism is legalVideo generationWindow: 2025-12-02 to 2026-10-18
Numbers are list positions, not rankings. Select a tool to jump to its profile. Every claim below comes from the sources in the table at the end, and each tool's window is the period its sources cover.
What do the signals say overall?
Users judge these tools less on raw capability than on three things that change after they have paid: how usage is metered, whether the answer's sources hold up, and where their data goes. Billing changes drew complaints for Cursor, ChatGPT and GitHub Copilot. Source accuracy is the main warning for Perplexity and Consensus. Privacy and ownership questions surfaced for Claude, ChatGPT and Cursor. For Midjourney and Runway we found almost no first-hand product feedback, only legal and strategy news.
AI tool pricing complaints: Cursor, ChatGPT and GitHub Copilot
The sharpest pattern is complaints about how usage is metered, not about the list price.
GitHub Copilot. GitHub says all Copilot plans moved to usage-based billing with AI Credits on 1 June 2026. Plan prices are unchanged, code completions stay included, and annual Pro and Pro+ subscribers stay on the earlier scheme until their plans expire (GitHub, 27 April 2026). Commenters report credits used up faster and a higher effective cost for the same work. Others say subsidised inference could not continue, which matches GitHub's stated reason that agentic sessions need far more compute.
ChatGPT. Many commenters called the Pro plan change a downgrade, because the $200 tier's included usage was reduced while its price stayed the same, and they criticised OpenAI for not stating concrete limits. A third-party report adds a $500 tier and says existing $200 subscribers keep the earlier allowance until 29 October 2026 (that date has passed, and we could not read OpenAI's own pages to check what applies now).
Cursor. Users complained that the usage page showed tokens instead of dollars. Cursor's pricing page, read on 2026-10-18, describes a usage pool shared between Cursor and Grok that resets weekly, with third-party models billed separately at API list price.
Claude. Commenters said heavy, token-hungry use is expensive and that employers are now setting budgets on AI tools; they disagreed on whether the top model pays for itself.
AI search citation accuracy: Perplexity and Consensus
Both tools are built around linking an answer to papers or pages, and the evidence says the links need checking.
Perplexity. A third-party audit by Haus Research ran 310 factual questions through Perplexity's API. It found that 34.7% of citations on sentences containing a figure pointed to a page that would not open or did not contain the figure (14.4% when scored per claim). The audit did not disclose its funding and tested the API, not the consumer app, and Hacker News commenters split on its credibility. An earlier Tow Center study (March 2025) found Perplexity answered 37% of news-source identification queries incorrectly, the lowest error rate of the eight tools tested but still large on a narrow task. The two studies used different tasks, so the figures cannot be compared.
Consensus. Library evaluations from Deakin, CASRAI and others recommend it for early scoping. They also say its Consensus Meter counts votes across studies without weighing study design or effect size, that for paywalled papers it mostly analyses titles and abstracts, that results are not reproducible enough for systematic reviews, and that it is weaker outside science and health. The vendor's own pages disagree on corpus size (200 million to 400 million or more), so we state no figure.
AI tool privacy concerns: Claude, ChatGPT and Cursor
Claude. A Hacker News thread about Anthropic reportedly passing a user's chat to police drew many comments saying people wrongly assume AI chats are private (the underlying report was not read, and no individual is named here). Anthropic's privacy centre, last updated 18 August 2026, says commercial products such as Claude for Work and the API are excluded from training by default, apart from submitted feedback or an explicit opt-in. Consumer plans are covered by a separate article we did not read.
ChatGPT. Commenters on two threads about advertising raised privacy concerns, saying people share personal details with chatbots and that ad targeting raises the stakes. OpenAI's own pages were blocked to our fetch tool, so we confirmed no policy.
Cursor. A user reported that installing the iOS app switched them to a new privacy mode they could not reverse. A commenter who says they work at Cursor replied that cloud agents need the new mode and that the legacy mode remains for people who do not use them (their identity is not verified). After the SpaceX acquisition, some commenters raised concerns about ownership and data use; Cursor's announcement does not address privacy.
GitHub Copilot. GitHub's Copilot page, read on 2026-10-18, says Business and Enterprise data is not used for training, and that other access retains prompts and suggestions for 28 days.
Tool by tool: what users praise, complain about and disagree on
Claude
Claude drew praise for Sonnet 5.5's price-performance and complaints about reliability, safety filters and the cost of heavy use. Commenters on the launch said Sonnet 5.5 comes close to the larger Opus 5.5 on agentic coding, and independent developer Simon Willison said it is nearly as good as Opus 5.5 on some coding tasks and is the free-tier model on claude.ai. These are impressions and vendor-reported comparisons, not our tests.
Complaints. Commenters reported repeated errors during an outage and said the status page lagged what they saw. Several said cyber-related safeguards flagged or downgraded them on ordinary work such as reviewing old C code. Disagreements. Whether the safeguards are sensible protection or an over-cautious "nanny" approach; whether session limits bite (some hit them quickly, others had no problems on similar plans).
What changed. Anthropic's release notes list a rapid model cadence: Opus 4.7 (16 April), Opus 4.8 (28 May), Fable 5 (9 June), Sonnet 5 (30 June), Opus 5 (24 July), Fable 5.1 (1 September), Opus 5.5 (22 September), Sonnet 5.5 (28 September) and Haiku 5.5 (7 October). Cowork and chat were merged into one experience on 16 September 2026.
ChatGPT
ChatGPT's recent discussion is dominated by the Pro plan change, advertising and two new products, Sites and ChatGPT Work. This record has the thinnest evidence of the group: a window of about five weeks, and none of OpenAI's own pages could be read.
Praise. Some heavy agent-style users said Pro still costs far less than equivalent API use. Some commenters reported good first experiences with Sites, including one who built a working 3D viewer. Willison found ChatGPT Work could assemble accurate boundary files from government data (one task, one person).
Complaints. The Pro change; ads and tracking; and Sites, which some called "underwhelming" or slow next to Claude's artifacts and a walled garden meant to lock in users. Willison called ChatGPT Work "an extraordinarily confusing and very powerful product". Disagreements. Whether ads on free tiers are acceptable, and whether a higher usage multiplier at the same price is a downgrade.
Cursor
Cursor's users split between those who still treat it as their main tool and those who have moved to Claude Code or Codex. The signals cover Cursor 3 (2 April 2026), the SpaceX acquisition (announced by Cursor on 14 August 2026), billing and Composer models.
Praise. Some developers value its Tab completions and staying in control of the code. One hands-on comparison by an engineering manager at a software agency credits reviewable diffs, isolated work environments and team controls such as privacy mode, single sign-on and audit logs. Complaints. Several commenters dislike the agent-first direction of Cursor 3 and want the editor to stay central; billing visibility and rising or hard-to-predict costs; frequent interface changes and tightening limits.
Disagreements. How good Composer is: Cursor's own benchmark reports strong results at low cost, while commenters cite an independent tracker that places it further behind. Whether the usage-page change was deliberate: Cursor staff on the forum call the removal of dollar figures deliberate, while a commenter who says they work at Cursor calls broken dollar values in the CSV export accidental and fixed. The same hands-on comparison's author prefers Claude Code personally.
GitHub Copilot
GitHub Copilot's discussion is dominated by the June 2026 move to usage-based billing, so sentiment may reflect that event more than product quality. GitHub also paused new sign-ups for Pro, Pro+ and Student plans on 20 April 2026 and tightened limits and model access.
Praise. Some commenters like Copilot CLI for feeling low-friction and for switching between models from different providers; some see Copilot as well placed for enterprises with several providers, hosted agents and CI integration. A preprint on Microsoft's internal rollout of command-line agents reports adopters merged about 24% more pull requests than expected over four months, and its authors caution that merged pull requests are an imperfect proxy for value.
Complaints. Billing, losing Opus access on Pro, claims that Copilot has fallen behind Claude Code and Codex, and the editor automatically listing Copilot as a commit co-author after inline suggestions. A security vendor, Wiz, reports a flaw that went live in a pull request that Copilot Autofix had co-authored and reviewed, at Snowflake; Wiz says Copilot did not write the vulnerable code, and the account rests mostly on Wiz's own claims.
Perplexity
Perplexity is credited for visible citations and live-web retrieval, and criticised because the citations do not always support the sentence they are attached to. Library guides recommend it for early scoping of an unfamiliar topic and advise treating its citations as "a lead, not a fact".
Complaints. The Haus audit and the Tow Center study above; guides say it is a general answer engine, not an academic search tool, and that live-web results are hard to reproduce exactly. Crawler conduct is questioned: the Tow Center authors suggest it may have disregarded publisher robots.txt preferences, one commenter's unverified report says AI crawlers including its own ignore robots.txt, and UBC's wiki notes copyright suits from media organisations.
Disagreements. Whether the Haus audit is credible (roughly 20 commenters took it at face value, about 10 questioned it), and whether citation mismatch is specific to Perplexity or common to AI search. A separate thread about whether a user-directed agent such as its Comet browser may act on sites like Amazon was mostly about law, not answer quality.
Consensus
Consensus is recommended by university libraries for early-stage scoping and warned against for anything that needs evidence synthesis. Yale is running a year-long trial through 31 December 2026, which signals institutional interest, not endorsement.
Praise. Library evaluations credit it for quickly surfacing relevant studies, linking back to papers and needing little training; reviewers like its filters, reference-manager export and full-text links. In one librarian's tests it declined a question with no relevant literature instead of inventing citations.
Complaints. The Consensus Meter's vote-counting, reliance on abstracts for paywalled papers, lack of reproducibility, and weaker coverage outside science, technology and health. Deakin also says vendor information on data sources, model training and retention is limited and flags copyright questions for the "Ask Paper" PDF upload.
Disagreement. How far its output can serve as evidence: library guides say it should not replace structured searching or reading the papers, while one Hacker News commenter shared a Consensus summary as supporting evidence against a mouse study. Community evidence here is five Hacker News comments, mostly single anecdotes.
Midjourney
For Midjourney we found little first-hand feedback on the image generator itself. The most-discussed item in the window was its medical-imaging announcement of June 2026.
Complaints. The main reaction in the large Hacker News thread was scepticism about the medical-imaging pivot: doubts about the scale claimed, the quality of the example images and the lack of published evidence, plus questions about overdiagnosis, false positives, liability and privacy. Some commenters said the company's reputation for AI-generated "slop" made them trust its medical claims less, which is a view about reputation, not image quality. Ben Recht's newsletter essay calls the ultrasound project mostly vaporware. Praise. Midjourney reports that early tests of a "Thinking" mode improve prompt accuracy, typography and coherence; that is a vendor claim, not independent evidence.
What changed. Five updates between 24 September and 8 October 2026, including two Alpha changelogs; a relayed but unconfirmed claim that v8.2 became the default model; and, per TechCrunch citing Bloomberg, the acquisition of the astrology app Co-Star in July 2026 (terms undisclosed).
Runway
For Runway we found no first-hand creator reviews we could open and verify, so this profile rests on vendor pages and legal reporting. Organic Hacker News discussion of its video product was thin in the window.
Praise. At launch in December 2025, Gen-4.5 was reported to rank first on Artificial Analysis' Video Arena leaderboard, which uses blind human comparisons. That measures preference, is dated, and we did not check the current standing. Complaints. The documented criticism is legal: three proposed or filed cases allege Runway bypassed YouTube protections to download videos for AI training. These are unproven allegations, and no ruling was reported.
What changed. Runway presents itself as a platform, not a single model: its homepage lists Creative, Dev (an API) and Robotics products, and its pricing page lists other labs' models (Kling, Seedance, Wan and Nano Banana Pro) alongside its own. These are the vendor's descriptions.
Limits of this guide
These are signals, not test results. Most community evidence is Hacker News, which skews toward developers, and most comments are single anecdotes, so we give no percentages beyond figures quoted from studies. Several records have low confidence: ChatGPT rests on no OpenAI page, and Midjourney and Runway have few sources. Reddit, Discord and review sites were not read.
Sources & Further Reading
| Claim | Source | Date |
|---|---|---|
| Claude: community discussion | Hacker News: Sonnet 5.5 launch thread. About eight top-level comments were readable in the part of the thread we read; the thread is much longer. | 2026-09-28 |
| Claude: community discussion | Hacker News: Claude partial outage thread. About 200 comments; we read a summary of the visible part. | 2026-09-29 |
| Claude: community discussion | Hacker News: Meta and Microsoft reduce employee use of Claude. About 130 comments seen. The headline claim comes from a news article we did not read. | 2026-10-07 |
| Claude: community discussion | Hacker News: Anthropic reportedly passed a user's chat content to police. About 80 comments seen. We did not read the underlying news report and name no individual. | 2026-10-05 |
| Claude: expert or practitioner view | Simon Willison: Claude Sonnet 5.5. Read through an index-page summary, not the full post. | 2026-09-28 |
| Claude: expert or practitioner view | Simon Willison: Tree Appreciator. A practitioner's demo of Opus 5.5 on one task; read through an index-page summary. | 2026-10-14 |
| Claude: vendor statement | Claude apps release notes (Anthropic). Date is that of the newest entry. Vendor statements about its own releases. | 2026-10-07 |
| Claude: vendor statement | How do usage and length limits work? (Anthropic). The page shows 'Updated this week', not a date; 2026-10-18 is when we read it. | 2026-10-18 |
| Claude: vendor statement | What are artifacts? (Anthropic). The page shows only a relative update time; 2026-10-18 is when we read it. | 2026-10-18 |
| Claude: vendor statement | Is my data used for model training? (Anthropic privacy centre, commercial products). Page states 'Last updated: August 18, 2026'. | 2026-08-18 |
| Claude: vendor statement | Commercial Terms of Service (Anthropic). Effective date shown on the page. Older than the six-month window but it is the governing contract text. | 2025-06-17 |
| ChatGPT: community discussion | Hacker News: Sites in ChatGPT. About 20 top-level comments were readable; the thread is longer. | 2026-10-01 |
| ChatGPT: community discussion | Hacker News: ChatGPT Pro 500. About 100 comments seen. The linked OpenAI help page returned 403 to us. | 2026-09-29 |
| ChatGPT: community discussion | Hacker News: ChatGPT and an ad collector. About 25 relevant comments seen. We did not read the linked blog post that made the claim. | 2026-09-20 |
| ChatGPT: community discussion | Hacker News: OpenAI expands ChatGPT ads with Sponsored Agents. About 150 comments seen. We could not read OpenAI's own announcement. | 2026-09-16 |
| ChatGPT: news report | WinBuzzer: OpenAI adds a $500 ChatGPT Pro plan and cuts the allowance for new $200 subscribers. Third-party report that quotes an OpenAI product leader and OpenAI's plan terms; we did not check OpenAI's pages. | 2026-09-30 |
| ChatGPT: expert or practitioner view | Simon Willison: Understanding ChatGPT Work. Read through an index-page summary, not the full post. | 2026-08-30 |
| ChatGPT: expert or practitioner view | Simon Willison: GeoJSON Map Viewer. One practitioner's experience on one task; read through an index-page summary. | 2026-09-01 |
| Cursor: community discussion | Hacker News: Cursor 3. Developer reaction to the agent-centered redesign; 544 points and 406 comments at collection. | 2026-04-02 |
| Cursor: vendor statement | Cursor: Cursor 3 announcement. Vendor description of the new interface and of the option to switch back to the classic IDE. | 2026-04-02 |
| Cursor: community discussion | Hacker News: Cursor Introduces Composer 2.5. Reaction to Cursor's own model and its cost claims; commenters are split. | 2026-05-18 |
| Cursor: community discussion | Hacker News: SpaceX to buy Cursor. Large thread (over 1,000 points) on the acquisition report and on whether developers still use Cursor. | 2026-06-16 |
| Cursor: community discussion | Hacker News: Installing Cursor on iOS changes your privacy settings. A user report; a commenter who says they work at Cursor replied (identity not verified by us). | 2026-06-30 |
| Cursor: community discussion | Hacker News: CursorBench 3.1. Commenters question the vendor's own benchmark against independent trackers. | 2026-07-02 |
| Cursor: community discussion | Cursor forum: usage page shows tokens instead of dollars. User complaints about billing visibility, with replies from Cursor staff on the same days. | 2026-07-31 |
| Cursor: community discussion | Hacker News: Cursor removed cost information from the usage page. Developer discussion of the change and of Cursor's value against Claude Code and Codex. | 2026-08-01 |
| Cursor: vendor statement | Cursor: Cursor is now a part of SpaceX. Vendor statement of the acquisition. The post makes no statement about privacy or data handling. | 2026-08-14 |
| Cursor: community discussion | Hacker News: Cursor is now a part of SpaceX. Smaller thread; commenters question the deal's rationale. | 2026-08-14 |
| Cursor: community discussion | Hacker News: Cursor launches Origin, a GitHub alternative. Reaction to Cursor's code-hosting product, including concerns about its ownership. | 2026-08-17 |
| Cursor: expert or practitioner view | DECODE: Cursor vs. Claude Code, in-depth comparison for dev teams. Hands-on comparison by an engineering manager at a software agency (published 12 Aug 2026, updated October 2026). The agency sells AI consulting, though it does not appear to sell either tool. | 2026-08-12 |
| Cursor: vendor statement | Cursor: pricing page. Undated vendor page read on 2026-10-18; used for how usage and privacy mode are described. | 2026-10-18 |
| Cursor: vendor release history | Cursor: changelog. First changelog page lists five entries from 10 Sep to 16 Oct 2026; earlier pages were not read. | 2026-10-16 |
| GitHub Copilot: vendor statement | GitHub: Changes to GitHub Copilot individual plans. Vendor announcement (updated 14 May 2026): paused sign-ups for some individual plans, tighter limits, changes to model access and refunds. | 2026-04-20 |
| GitHub Copilot: community discussion | Hacker News: Changes to GitHub Copilot individual plans. Developer reaction; many comments focus on model access and pricing. | 2026-04-20 |
| GitHub Copilot: vendor statement | GitHub: Copilot is moving to usage-based billing. Vendor explanation of the 1 June 2026 billing change and of what stays included. | 2026-04-27 |
| GitHub Copilot: community discussion | Hacker News: GitHub Copilot is moving to usage-based billing. Very large thread (over 700 points) on the billing change. | 2026-04-27 |
| GitHub Copilot: community discussion | Hacker News: VS Code adds GitHub Copilot as a co-author. Users report the editor adding Copilot as a commit co-author after inline suggestions were accepted. | 2026-04-30 |
| GitHub Copilot: news report | The Register: Angry devs vow to flee GitHub Copilot as metered billing takes hold. Trade-press report quoting forum and Reddit users and a GitHub spokesperson; the cost figures are users' own and not verified by the outlet. | 2026-06-02 |
| GitHub Copilot: community discussion | Hacker News: Angry devs vow to flee GitHub Copilot. Smaller thread reacting to the Register piece. | 2026-06-02 |
| GitHub Copilot: community discussion | Hacker News: GitHub Copilot App. Reaction to GitHub's desktop app for working with agents. | 2026-06-02 |
| GitHub Copilot: expert or practitioner view | arXiv: Adoption and Impact of Command-Line AI Coding Agents (Murphy-Hill, Butler, Savelieva). Preprint on Microsoft's early-2026 rollout of Claude Code and Copilot CLI. It measures merged pull requests, which the authors say do not equal delivered value. | 2026-07-01 |
| GitHub Copilot: community discussion | Hacker News: Kimi K2.7 Code is generally available in GitHub Copilot. Thread links GitHub's changelog; commenters discuss Copilot's model choice, CLI and pricing. | 2026-07-02 |
| GitHub Copilot: community discussion | Hacker News: A Study of Microsoft's Early 2026 Rollout. Commenters question what merged pull requests show. | 2026-07-13 |
| GitHub Copilot: expert or practitioner view | Wiz: Red Agent finds its way into Snowflake's Jira (Copilot Autofix). Security-vendor account. Wiz sells security products, and most of its technical claims are its own; Snowflake's published statement corroborates only the fix and the credential rotation. | 2026-08-17 |
| GitHub Copilot: community discussion | Hacker News: AI-Generated GitHub Copilot Autofix allowed compromise of Snowflake's Jira. Developer discussion focused on CI security practice. | 2026-08-17 |
| GitHub Copilot: vendor statement | GitHub: Copilot product page. Undated vendor page read on 2026-10-18; used for plans, supported IDEs, agents and data-retention wording. | 2026-10-18 |
| Perplexity: expert or practitioner view | Haus Research: Perplexity citation audit (HR-2026-09). Third-party audit of 310 factual questions run through the API, not the consumer app; the page discloses no funding or conflicts, and its own limitations include a single snapshot and hand-picked companies. | 2026-09-02 |
| Perplexity: community discussion | Hacker News discussion of the Haus audit (about 48 comments). Forum skews toward developers; counts in our summary are approximate. | 2026-09-02 |
| Perplexity: expert or practitioner view | Columbia Journalism Review, Tow Center: "AI search has a citation problem". Tested identifying the source of news excerpts, which is a narrow task; the authors say the findings should not be extrapolated to all models or publishers. Older than the rest of the window. | 2025-03-06 |
| Perplexity: expert or practitioner view | CASRAI Editorial Board: Perplexity AI for research. Guide from a programme of a US nonprofit; it says vendor claims about search passes were not independently verified. | 2026-08-31 |
| Perplexity: expert or practitioner view | UBC Library Wiki: Perplexity. Last-edited date of a librarian-compiled wiki entry; it cites several studies by reference rather than testing the tool itself. | 2026-09-28 |
| Perplexity: community discussion | Hacker News discussion: Amazon v. Perplexity, Ninth Circuit. Mostly about the law of browser agents, not Perplexity answers; our counts come from the first part of a long thread. | 2026-09-14 |
| Perplexity: community discussion | Hacker News comment: daily paid use. A single anecdote. | 2026-09-30 |
| Perplexity: community discussion | Hacker News comment: answers and citations vary between runs. A single comment from someone building a brand-tracking tool, who has a stake in the point. | 2026-07-12 |
| Perplexity: community discussion | Hacker News comment: crawler behavior report. An unverified first-person report from one site operator. | 2026-10-18 |
| Perplexity: vendor statement | Perplexity blog (recent posts). Date is the newest post listed when we read the page; the vendor describes its own direction. | 2026-10-16 |
| Perplexity: vendor statement | Perplexity Help Center: What is Perplexity?. The page is undated; this is the date we read it. | 2026-10-18 |
| Perplexity: vendor statement | Perplexity Help Center: Spaces (no link: the page was read on 2026-10-18 but returned 404 when we rechecked on 2026-11-06). The page was undated. | 2026-10-18 |
| Perplexity: community discussion | Hacker News: Show HN, CiteRank (tracks whether AI tools cite your site). We saw the title and metadata through the HN search API and did not read the thread. | 2026-09-25 |
| Consensus: expert or practitioner view | Deakin Library: AI evaluation of Consensus. Library evaluation, first published June 2025 and last reviewed August 2026; month only, no named author. | 2026-08 |
| Consensus: expert or practitioner view | CASRAI Editorial Board: Consensus AI search engine guide. Guide from a programme of a US nonprofit; a reviewed, general-audience assessment. | 2026-08-24 |
| Consensus: news report | Yale University Library: year-long trial of Consensus. Announces a trial through 31 December 2026; it repeats the vendor's promise of evidence-backed answers and lists no limitations. | 2026-03-20 |
| Consensus: expert or practitioner view | UBC Library Wiki: Consensus. Last-edited date of a librarian-compiled entry that summarizes other reviews. | 2026-07-17 |
| Consensus: expert or practitioner view | Aaron Tay (librarian): A 2025 deep dive of Consensus. Hands-on review by a librarian writing on a personal newsletter; older than the 2026 sources. | 2025-11-15 |
| Consensus: expert or practitioner view | HKUST Library: Trust in AI, evaluating Scite, Elicit, Consensus and Scopus AI. One-prompt comparison from 2024; the tool has changed since, so treat it as historical. | 2024-03-20 |
| Consensus: vendor statement | Consensus blog: Deep Search. Vendor description at launch of a beta feature; the page metadata shows a later date. | 2025-07-09 |
| Consensus: vendor statement | Consensus homepage. The page is undated; this is the date we read it. | 2026-10-18 |
| Consensus: community discussion | Hacker News comment: preparing for a doctor visit. A single anecdote. | 2026-08-06 |
| Consensus: community discussion | Hacker News comment: starting point for reviewing treatment research. A single anecdote. | 2026-03-31 |
| Consensus: community discussion | Hacker News comment: listed among AI literature-search competitors. Mention in a discussion of another product; no evaluation of Consensus. | 2026-01-27 |
| Consensus: community discussion | Hacker News comment: listed among paper-mapping tools. Mention in a Show HN thread; no evaluation of Consensus. | 2026-07-09 |
| Consensus: community discussion | Hacker News comment: Consensus summary shared as supporting evidence. A single comment. | 2025-10-09 |
| Midjourney: community discussion | Hacker News thread on Midjourney Medical. Large thread (over 100 comments read; the full thread is far longer). It is about the medical product, not image generation. | 2026-06-18 |
| Midjourney: expert or practitioner view | arg min: "All data is good data". Newsletter essay by Ben Recht on the Midjourney ultrasound announcement. Opinion. | 2026-06-22 |
| Midjourney: news report | TechCrunch: Midjourney acquired the astrology app Co-Star. Brief that cites Bloomberg for the deal. Deal terms undisclosed. | 2026-07-24 |
| Midjourney: community discussion | Hacker News: "Midjourney Releasing v8.2". Two-point submission with no link and two short comments; it relays a release claim we could not confirm from Midjourney. | 2026-07-25 |
| Midjourney: vendor statement | Midjourney updates: Testing Thinking Mode. Vendor announcement. Its statements about improved accuracy are Midjourney's own early-test claims. | 2026-10-08 |
| Midjourney: vendor statement | Midjourney updates index. Lists five posts between 24 September and 8 October 2026, including Alpha changelogs. | 2026-10-08 |
| Midjourney: news report | Graphic Policy: studios and Midjourney discovery dispute. Reports filings in the studios' copyright case against Midjourney. It reports no ruling and does not describe the substance of the dispute. | 2026-05-11 |
| Runway: vendor statement | Runway homepage. Undated page read on the collection date. All product and partnership statements are Runway's own claims. | 2026-10-18 |
| Runway: vendor statement | Runway pricing page. Undated page read on the collection date. Plans and model lists change often. | 2026-10-18 |
| Runway: news report | RightsTech: YouTuber sues Runway AI. Reports a proposed class action filed in Los Angeles. The article reports no response from Runway. Allegations only. | 2026-02-25 |
| Runway: news report | ChatGPT Is Eating the World: Runway sued in a fourth copyright case. Independent legal tracker. It counts four cases and reports three with similar YouTube-related DMCA theories. No rulings reported. | 2026-03-02 |
| Runway: news report | eWeek: Runway launches Gen-4.5. Launch coverage that relays Runway's claims and an Artificial Analysis leaderboard result. It includes no independent testing. This predates most of the window. | 2025-12-02 |
| Runway: community discussion | Hacker News search: Runway stories. Search listing. The Runway-related submissions found are mostly Runway's own engineering blog posts. Date shown is the earliest such post. | 2026-04-27 |
Frequently asked questions
What do users say about AI tools in 2026?
Public discussion of eight AI tools, read in October 2026, centres on cost changes, source accuracy and privacy more than raw capability. Billing drew complaints for Cursor, ChatGPT and GitHub Copilot; Perplexity and Consensus draw citation warnings. These are signals, not test results.
Why are users unhappy with GitHub Copilot pricing?
GitHub moved all Copilot plans to usage-based billing with AI Credits on 1 June 2026. Plan prices stayed the same, but commenters report credits running out faster and a higher effective cost. Others say subsidised inference could not continue. Annual subscribers stay on the earlier scheme until expiry.
Can you trust Perplexity citations?
Treat them as leads. A third-party audit of 310 questions found 34.7% of citations on figure-bearing sentences pointed to a page that would not open or lacked the figure (14.4% per claim). Its funding was undisclosed and it tested the API. Library guides advise checking the primary source.
Is Consensus good for systematic reviews?
Library guides say no. They recommend it for early scoping, but note its Consensus Meter counts votes without weighing study quality, it relies on abstracts for paywalled papers, and results are not reproducible enough for systematic reviews. Check the linked papers before relying on a summary.
How we sourced this. Signal records collected 2026-10-18 from the sources below; our editors did not re-read them for this page. We tested none of these tools. Methodology






