Where does Grok 4.5 fall short of

@techdevnotes limits?

@techdevnotes Cross chat memory, still the shittiest UX of any frontier models. Sure it can hit benchmarks who cares when it’s got a crap harness

@techdevnotes People are hitting limits really fast, so I would rather stick with Claude Max, which works great for me. Until SpaceX releases a transparent plan, I am not buying it, the sub models are sketchy. @elonmusk make a generous plan for coding only without the image/video nonsense.

@techdevnotes Usage limits certainly make it less usable

@techdevnotes The limits are so miniscule it's genuinely unusable. Also feels like the reasoning itself kinda went in the shitter. Having a decent convo feels empty and repetitive.

@techdevnotes running a pilot for work.. I’m impressed so far but will post if something comes up!

Your first Gopuff order gets $0 delivery, plus 100s of deals once you're in. 5,000+ products, no markups, delivered in as fast as 15 minutes.

@techdevnotes When they first launched Grok 4.5 last week, its really efficient usage wise and agentic tasks, but now they silently nerfed it. I noticed it. Ive been using it since the day they launched it.

@techdevnotes I think he is not thinking deep enough, and sometimes he don't want to ask user for clarification.

@techdevnotes Limits and generation in blender it cannot for whatever reason properly setup blender characters

@techdevnotes Weekly limits is insane, telling my companion about my shit day spills not be taking much compute

@techdevnotes Not the model itself, but Usage limits and the clarity around them. I ran out quickly with a SuperGrok sub, but i have no idea what I would get with a Heavy sub and if it justifies the cost.

@techdevnotes It’s very clear by all the comments that it falls short with the usage limits.

@techdevnotes in usage and limits... its ridiculously low! Claude feels better. there's gotta be a bug or something...

Well, sometimes, when I’ve asked Grok Imagine to use a certain art style in a creation, it ignores those requests. & as far as queries go, if I don’t go back to the same original query & add the new question @ the same topic to that, but start a new query over the same past info, Grok doesn’t remember all the info from the other query

@techdevnotes Making its way through eu bureaucracy 🤣

@techdevnotes Moderation remains unintelligent and overly restrictive on obvious non-realistic characters. This is one area Grok has consistently struggled with.

@techdevnotes Even Cheaper, knock out the chinese models out.

@techdevnotes the fact it eats up 2% of my weekly use per small feature

@techdevnotes Same places as before because they didn't change the human element which is overmoderating image/video creation to an absurd degree. Limits are still a problem but the moderation issues affect that too.

Up late and bored? Tap into live conversations.

@techdevnotes Limits are my biggest complaint currently. They feel much better on Chatgpt Pro and even Claude felt better on their 20x plan.

@techdevnotes Memory and context make it poor for big projects unless you want to kind of baby sit it.

@techdevnotes Why isn’t it in the Web or iOS app?

@techdevnotes Not being available in the app? Like at all???

@techdevnotes At least in real-world agentic performance: Grok 4.5 currently ranks just 13th on Agent Arena—even below GPT-5.4. Strong benchmark scores don’t always translate into reliable long-horizon execution. arena.ai/leaderboard/ag…

@techdevnotes For website work, fantastic, I had it build me a Wispr Flow alternative and lots of bugs, but after much iteration the app is working great now.

@techdevnotes Where does X fall short? Customer support. Take a minute, go to Reddit and read of all the accounts suspended from X unjustly and how there’s no support to contact.

@techdevnotes @SpaceXAI literally community feedback in this whole comment section

@techdevnotes Grok high burns through limits too fast

@techdevnotes UX - being able to drop screen shots of configs, errors, etc. is faster/easier in Claude. To be fair, with the latest release, I’ve been getting equivalent results without the screen shots, so maybe I don’t need it.

@techdevnotes Usage limits. I want to switch from ChatGPT Pro but looks like I'll hit the limits with SuperGrok Heavy.

@techdevnotes Uploading your data to its servers x.com/xbtoshi/status…

@techdevnotes It falls short on something known as a benchmark

@techdevnotes в подписке за 30$ расход лимита за 2 дня и потом ждать еще 5 дней отката

@techdevnotes Astrology, in my experience, fable is miles better sadly

@techdevnotes Limits are heinous.