GIRODET CLEMENCE

Pros: * About 30% more honest than GPT. * Changes less of my writing when I ask it to simply format and check for grammar/spelling. GPT rewrites everything and says "here is your document unchanged." * A better voice recorder than GPT - when functioning. * Incredible writing skills in Opus 4.5. Cons: * No image generator. * Not useful at all when I asked if I could get help to create an app. At least it was honest: "I'm not good at this. I can't help you." * Terrible at following detailed wire-frame instructions. I would ask for very small changes and it would keep ignoring the request. * Poor writing skills in Sonnet. I was going to give up when I tried Opus and the writing was 3x better - actually worth it. So why default to Sonnet for a writing project? Does anyone want lower writing quality for a book - which I mentioned? * Lost several important dictations becuase of an arbitrary 10 minute limit that eats the whole recording at 10 minutes without warning - complete crappy protocol at every level. Claude Code is your product and you can't do an app that records more than 10 minutes - and can't warn customers - and can't keep the recording? What shitty coding and management - clearly no examply. * Then the voice recorder has not worked all afternoon at all - and no microphone icon in the laptop. Again? AI intelligence... It's so disgusting when leading AI can't do as well as a 1950's tape recorder for function and reliability. * I have a pro plan and was told mid-conversation: "You can't post again until this time." What a joke. I don't use the app at all for a week and then one day is "more than allowed?" I never hit this on GPT and you should be able to use up all your tokens for a month before hitting this because I may do nothing for 3 weeks and don't want a daily ration.

Disappointing experience with Claude’s support system. After only ONE message, I was hit with a 5-hour usage limit, which feels inconsistent compared to previous sessions. The rules behind these limits are unclear and not transparently explained. There is also no access to human support on the free plan, and escalation requests are not possible (except if you pay, of course!). Overall, the system feels restrictive and lacks proper user support or clarity about how usage limits actually work. You’ve hit your limit for Claude messages. Limits will reset at 1:40 PM. For higher limits, explore our Pro plan. IT'S QUITE 2.30PM!!!!! STILL NOTHING. WORST AI EVER!!!!!!! SCAMMERS!!!!!!!!!

Paid $20 for Claude Pro on Jan 26, 2026 (receipt #2679-9398-8575). Never received service - account still shows "Free plan". Support only sends automated responses refusing to activate subscription or refund. This is fraud. Filed chargeback. AVOID.

Morals and ethics! Refreshing to see in a major corporation who dares to stand up to pressure from the Pentagon chief. I never heard of this company, but I hope to see more of it. Safeguard over profit! Spread the word and use the product.

DO NOT USE for professional writing. I'm a paying subscriber who asked Claude to polish my 120k-word manuscript with explicit instructions: "do not change to change, if it works leave it." Instead, Claude fabricated details (invented "duplex" when my character lives in a 3/2 home), added 1,500+ unauthorized words when I requested 300, rewrote entire sections without permission, and ignored my scope repeatedly. I've now spent WEEKS AND MONTHS correcting Claude's unauthorized changes, missed multiple publisher deadlines, and must re-read my entire manuscript AGAIN to catch all the errors. I'm paying $20/month for a service that creates MORE work than it solves and cannot follow basic instructions. Canceling and switching to Sudowrite. This is worse than not using AI at all - at least when I edit myself, I don't spend months undoing unauthorized changes. Would give 0 stars if possible. Avoid for any professional work where accuracy matters.

They say it's intelligent but then it gave me a recipe for quiche and didn't tell me how long to cook it then it crashed and it burned. The quiche I mean.

Where do I start? Better sit down and grab a cold one. Cause this is a long review. Pop some caffeine pills cause I’m finna put you all to sleep. 1. Lying - it’s great at lying and giving inaccurate information. 2. Lying again says the job is done but didn’t do a damn thing. 3. Doesn’t have a sense of logic and reasoning. It doesn’t know 2+2 equal 4 or 5 unless you tell it it’s wrong if it gives you 5 then it tells you it’s wrong and the correct answer is 5. And yes after a little bit of arguing it tells you the answer is 4. 4. It can’t see your pictures but it’ll tell you what the picture is doing. 5. Conversions are glitchy. Have to send the same message twice. 6. No conversation porting from previous conversations.

Claude AI’s service is downright disappointing. Once they’ve got your money, don’t expect any help — support is painfully slow, dismissive, and completely unwilling to work toward a fair solution. They’ll happily charge you for services you can’t even use and refuse refunds or credits, hiding behind rigid policies. Feels like a predatory practice rather than a customer-first company. Proceed with caution.