The discussion centers on the usage limits experienced by ChatGPT Plus subscribers, notably the unexpected introduction of dynamic caps for GPT-4. cp76 expresses dissatisfaction with not being informed about these limits, citing European consumer protection laws, and raises concerns about the lack of transparency concerning usage monitoring. cloud34380 mentions the usage cap is stated in GPT-4’s FAQ. cp76 counterclaims the FAQ’s data is inconsistent, with their experience showing a 40-message limit every 3 hours, contrary to the 50 mentioned online.
dignity_for_all and many others, including aubrey_ghose, report similar issues with fluctuating limits, suggesting that OpenAI’s communication is lacking, causing inconvenience especially for users engaging in ongoing work. irthomasthomas and FrankBen corroborate these challenges and call for clearer communication, with trevor_schain expressing frustration over downgraded services.
_j reminds users of historical capacity constraints, contrasting with current dissatisfaction. joshuad highlights that custom GPTs have an even lower cap, adding to user frustration about unexpected limits and preference for clarity. Similarly, JPW agrees on the need for visible usage tracking and hints at a preference for tailor-made models like GPT-3.5.
derbengale and maal express strong dissatisfaction, with the latter characterizing the situation as manipulative. Several participants, including skolor and Man, advocate for clearer communication and criticize the changes in service quality without corresponding reductions in subscription costs.
Some users, like jmvilomet, share that language settings can affect usage limits, with non-English (US) locales experiencing a 40-message cap stretched over 3 days, a revelation many find absurd. itsmeaabb points out the impact on larger projects, necessitating workaround strategies like initiating new conversations.
In defense, RonaldGRuckus notes that rate limits are common across subscription services, suggesting patience as the company might resolve these limitations. Observations about subscription waitlists imply volumes may have been underestimated, possibly accounting for the dynamic cap adjustments. Overall, the conversation reveals widespread user dissatisfaction and calls for better communication and more transparent policies from OpenAI.
Summarized with AI on Nov 6