What do users say about ChatGPT?

ChatGPT is great for brainstorming new ideas. It really feels like a personal AI assistant—almost like having a partner to bounce thoughts off of. Review collected by and hosted on G2.com.

I’ve noticed hallucinations, especially when I ask for recent tech news. It will claim the news is fresh, but when I open the source link, the article turns out to be about two months old or more. Review collected by and hosted on G2.com.

I like ChatGPT because this AI agent gives some of the best answers among other AI agents. It knows how to respond in different scenarios, and I also like that we can train ChatGPT so it can do what we want it to do as an agent, but in an ethical way. Review collected by and hosted on G2.com.

What I dislike about ChatGPT is that it often repeats the same answer when I ask a related question. At first it gives me the same solution again, and only after that does it provide a different one. That back-and-forth feels time-consuming and irritating, which is why I don’t like it. Review collected by and hosted on G2.com.

The best thing about ChatGPT for me is how good it is at breaking down messy patterns and helping me write quick regex rules or Python scripts to automate parts of my daily queue. Working in Trust and Safety on anti-phishing tasks, I run into a lot of weird obfuscated URLs, base64 strings, and sketchy redirect chains. Being able to paste a de-fanged URL structure or a block of suspicious HTML source code into ChatGPT and have it explain the exact social engineering trick or write a quick regex filter to catch similar phishing domains is a huge relief. The custom instructions feature is also super handy because I dont have to keep explaining my formatting rules every single time I start a new chat during a busy shift. Review collected by and hosted on G2.com.

The biggest annoyance for me is definitely the strict content filters getting triggered by false positives. Since my job involves reviewing anti-phishing queues and investigating scam websites, I'll sometimes drop in a screenshot of a fake login page or a suspicious URL string to analyze, and ChatGPT will just block the prompt and give me a generic safety warning even though I'm literally doing defensive moderation work. Also, when the servers get busy during peak hours, the image upload and web browsing features can lag quite a bit, or it will randomly make up domain history details instead of admitting it doesn't know a smaller regional brand. You still have to double check its outputs against actual WHOIS lookups so you don't make a bad call on a ticket. Review collected by and hosted on G2.com.

Using Chat-gpt is very easy; you just type the prompt you want and get an instant result. The onboarding is straightforward too—I signed up with my email and started working right away without any hassle. The intelligence of Chat-gpt feels smooth and reliable. I use it to generate code for training my ML models, edit my images, do research, and summarize that research into a PDF. Performance has been consistently smooth for me, and it usually returns results within minutes after I submit a prompt. Integrations are also well supported: I can upload files and images for my work, and I use it alongside Sheets, Docs, and similar tools. Pricing feels reasonable overall; I mostly stick to the free version, which is enough for my day-to-day tasks. Review collected by and hosted on G2.com.

Sometimes the performance is slow and it gives vague answers. The monthly plan also feels a bit expensive for an individual, especially for me as a student. At times it doesn’t understand my prompts well and ends up giving completely unclear responses, and it can even contradict its own answer in a follow-up prompt. For example, when I try to solve a maths question with Chat-gpt, it sometimes gives two different answers to the same question when I ask it in two different prompts. Review collected by and hosted on G2.com.

Strong conversational flow: It handles long back-and-forth threads very well, allowing you to gradually refine an idea or tone across multiple prompts without losing track of instructions.

Versatile creative brainstorming: It produces diverse angles, hooks, and creative concepts when exploring different ways to frame a topic or script.

Clear, step-by-step breakdowns: Great at explaining complicated topics in simple terms or breaking down a vague idea into actionable, ordered steps. Review collected by and hosted on G2.com.

Workflow & Ecosystem Friction

Formatting breaks during export: Copy-pasting structured content—especially tables, nested bullet points, or video scripts—into Google Docs frequently broke formatting and line spacing, forcing me to spend extra time fixing the layout manually.

Slower multimodal processing: Uploading screenshots, image references, or mockups took noticeably longer to parse and analyze, especially during peak traffic hours when image response times lagged.

Feels isolated from daily tools: Because it operates strictly in its own silo, moving drafts, meeting notes, or research into daily documents required constant manual downloading, re-uploading, and tab switching.

Occasional context slip in long chats: In longer brainstorming threads, it would sometimes forget negative constraints (like "don't use bullet points" or specific word limits) established earlier in the conversation. Review collected by and hosted on G2.com.

What I like most is how versatile it is. I use it throughout the day for debugging, writing, researching things, and figuring out problems I haven't dealt with before. Being able to give it a messy explanation, screenshots, or a bunch of context and get something genuinely useful back saves me a lot of time. The memory and ability to pick up conversations where I left off are also really helpful. Review collected by and hosted on G2.com.

Sometimes it's way too confident about things that are just wrong, which gets frustrating, especially when debugging. It also tends to overexplain simple questions or make writing sound overly polished and robotic. I often have to ask it to be more direct or rewrite things to sound natural. Memory is useful, but it doesn't always remember the right context. Review collected by and hosted on G2.com.

Chatgpt is commendable i mean to say it's a miracle.Its is beneficial to students, to corporates and people from all ages groups. Chatgpt is game changer for today's era . It's very time and effort convienient and stress reducer for people. The main is it's completely free and therefore people are using it the way they want. People are finding new things and dimishing old findings patterns and taking look at new place, new ways for something, new knowledge and vast generational change. Review collected by and hosted on G2.com.

Chatgpt is although a time saver but it's very cruel to students who don't want to work hard and learn things on there own..it's there age to learn and they are misusing it heavily without knowing it's disadvantages. People are searching any cheap things which are not good for society and for them. Some people are finding shortcuts to problems and other bad things which is not ignored. Review collected by and hosted on G2.com.

I use ChatGPT daily across game development tasks and general office administration. On the dev side, it works really well for quickly sanity checking logic, tracing syntax bugs, and writing boilerplate scripts without having to dig through outdated forum threads. On the office side, it cuts down on routine friction by taking rough meeting bullet points or requirements and shaping them into clear sprint updates and documentation for the team. Having a fast sounding board that handles repetitive drafting and error troubleshooting saves me significant context switching each week. Review collected by and hosted on G2.com.

The biggest issue is that it can be confidently wrong when dealing with newer framework versions or library updates. In game development, it frequently suggests deprecated API methods or invents functions that look completely plausible but do not actually exist in the engine, which leads to wasted time tracking down ghost errors. In longer conversation threads, it also tends to lose context or forget specific constraints mentioned earlier, requiring repeated re-prompting. On top of that, responses can be overly padded with generic advice and conversational filler when all I really need is a direct answer, a short diff, or a clean code snippet. Review collected by and hosted on G2.com.

The best thing I love about Chatgpt is that it's efficiency and accuracy. It usually responds quickly and most importantly the answers are correct. It's my go to AI platform, for me chatgpt has replaced google, everything I have to search, I search on chatgpt rather than Google or any search engine. Also, I like that it offers enough number of queries and attachment of images and documents in the free tier accounts. Review collected by and hosted on G2.com.

Certain shortcomings of chatgpt are, it lags somewhere in heavy tasks especially in coding, sometimes it doesn't give the most accurate code which I ask for. Another thing which I personally don't like is that it lags way too much when the chats get bigger. Also, image generation could be made better I think. Review collected by and hosted on G2.com.

ChatGPT noticeably speeds up my day-to-day developer workflow by acting as an on-demand sounding board for technical architecture decisions, code generation, and debugging. When I’m building integrations between Salesforce and AWS, it can draft boilerplate Apex classes, Lightning Web Components (LWC), and Node.js/Python serverless scripts with a high level of accuracy.

Beyond generating code, it’s also strong at parsing complex payload schemas, unpacking obscure platform error messages, and formatting JSON structures, which saves me hours of routine research and documentation work. It also cuts down on context switching by pulling what would normally require multiple documentation lookups into a single conversational interface. Review collected by and hosted on G2.com.

While its broad knowledge base is extensive, it can occasionally produce deprecated syntax or platform-specific methods that don’t meet modern runtime constraints—for example, referencing outdated Salesforce API versions or relying on legacy AWS SDK v2 patterns instead of v3.

Also, when I paste long technical specifications or large codebases, context retention can weaken over extended sessions. In those cases, it may drop constraints that were defined earlier or hallucinate parameters that aren’t actually present in the official API spec. Review collected by and hosted on G2.com.