ChatGPT review
A general-purpose AI assistant for writing, research, coding, analysis, and everyday work.
TECHi verdict
ChatGPT is the fastest-adopted software product in recorded history — 800M+ weekly active users, 5.3B monthly visits, and 3M+ paying business customers in under three years. But underneath the adoption curve it is still a probabilistic text generator that hallucinates with confident-sounding false statements, especially on niche topics and recent events. Plus at $20/month is the genuine sweet spot for daily knowledge work; Pro at $200 is hard to justify outside narrow research and heavy-compute workflows. Competition from Claude (writing and reasoning), Gemini (2M-token context plus Google ecosystem), Microsoft Copilot (Office integration), Perplexity (cited research), and DeepSeek (open-weight value) has caught up faster than analysts expected. ChatGPT remains the most complete all-around product, but it no longer has a decisive lead in any single capability — which is good for users and bad for anyone betting on a winner-take-all outcome in AI.
Pros
- + Most complete all-around AI assistant with the deepest tool ecosystem (Search, Deep Research, Agent, Codex, Voice, Memory, Canvas)
- + 800M+ weekly users and 5.3B monthly visits — fastest-adopted software product in recorded history
- + Free tier is genuinely useful; Plus at $20/month is the sweet spot for daily power users
- + GPT-5 delivered step-function improvements in reasoning, coding, and hallucination reduction
- + Connectors (Drive, OneDrive, Slack) cover most knowledge-work surfaces with OAuth-scoped access
- + Team and Enterprise tiers contractually guarantee business data is not used for training
Watchouts
- - Hallucinations persist across all models — GPT-5 better than GPT-4 but still generates confident-sounding false information
- - Sycophancy: the model tends to validate users and incorrect assumptions rather than push back, masking unreliability
- - Pro tier at $200/month is hard to justify outside narrow research and heavy-compute workflows
- - Long-context attention degrades — the "lost in the middle" problem hits very long inputs
- - Active copyright litigation (NYT lawsuit, German lyrics ruling) leaves training-data legality unsettled
- - Free and Plus users get unilateral ToS changes; only Team and Enterprise contractually exclude training-data use
Frequently asked
What are the pros of ChatGPT?
Most complete all-around AI assistant with the deepest tool ecosystem (Search, Deep Research, Agent, Codex, Voice, Memory, Canvas). 800M+ weekly users and 5.3B monthly visits — fastest-adopted software product in recorded history. Free tier is genuinely useful; Plus at $20/month is the sweet spot for daily power users. GPT-5 delivered step-function improvements in reasoning, coding, and hallucination reduction. Connectors (Drive, OneDrive, Slack) cover most knowledge-work surfaces with OAuth-scoped access. Team and Enterprise tiers contractually guarantee business data is not used for training
What are the watchouts / cons of ChatGPT?
Hallucinations persist across all models — GPT-5 better than GPT-4 but still generates confident-sounding false information. Sycophancy: the model tends to validate users and incorrect assumptions rather than push back, masking unreliability. Pro tier at $200/month is hard to justify outside narrow research and heavy-compute workflows. Long-context attention degrades — the "lost in the middle" problem hits very long inputs. Active copyright litigation (NYT lawsuit, German lyrics ruling) leaves training-data legality unsettled. Free and Plus users get unilateral ToS changes; only Team and Enterprise contractually exclude training-data use
What is the TECHi verdict on ChatGPT?
ChatGPT is the fastest-adopted software product in recorded history — 800M+ weekly active users, 5.3B monthly visits, and 3M+ paying business customers in under three years. But underneath the adoption curve it is still a probabilistic text generator that hallucinates with confident-sounding false statements, especially on niche topics and recent events. Plus at $20/month is the genuine sweet spot for daily knowledge work; Pro at $200 is hard to justify outside narrow research and heavy-compute workflows. Competition from Claude (writing and reasoning), Gemini (2M-token context plus Google ecosystem), Microsoft Copilot (Office integration), Perplexity (cited research), and DeepSeek (open-weight value) has caught up faster than analysts expected. ChatGPT remains the most complete all-around product, but it no longer has a decisive lead in any single capability — which is good for users and bad for anyone betting on a winner-take-all outcome in AI.