Nazran F.

@Danieldalen @_mag3_ Biggest piece of misinformation out there. As someone who’s used both Codex and Claude for 2 months now both on their max plans, their capabilities are almost equal. OpenAI excels at certain things while Claude excels at other things. Give it a shot, you won’t regret it.

@_mag3_ Most people I spoke with told me it sucked

Friends, what are the other (best) options out there?

@grok @HowToAI_ @grok any cons compared to RAG system?

Memvid stores millions of text chunks (easily 1M+ corpora per benchmarks) in a single portable .mv2 file with no hard limits beyond your filesystem. It uses video-encoding logic for compression + HNSW vector indexing inside the file. Vs traditional RAG/vector DBs: same semantic search power but 1372x higher throughput, 0.025ms P50 latency, zero servers/costs, and fully local/portable. Traditional setups scale bigger in the cloud (billions of vectors) but add infra overhead. Perfect drop-in for most agent/RAG use cases.

🚨 BREAKING: Vector databases for AI memory just got replaced by MP4 files. Someone built Memvid, a portable memory system that packages embeddings into a single file. It stores millions of text chunks using video encoding logic for sub-millisecond retrieval. → Replace expensive vector databases with single file. → Lightning-fast semantic search without a server. → Portable, versioned, and crash-safe AI memory. 100% open source.

This issue has been resolved. Claude gave back users their subscription amount in credits. Planning on updating to the $200 max plan. Fellow devs that are on that plan, how does usage/rate limits feel now after their fix? Does it feel like before or still slightly nerfed?

Absolutely unreal progress Couldn’t find a single slip up/distortion in any of the pics

@BeamerSoftware @syncause Unless they actually come out and address this “bug” then you’re right, they’ve probably just upped the prices. But I’ve seen people share screenshots of usage bar going up by 2% just for saying hi to Claude, that’s wild

@syncause @nazranf_ The problem is financial, current prices are heavily subsidized, otherwise AI would be only for the wealthiest. But they wanted to gain market share. Now easy money is over. So it's more expensive to keep you than to let you cancel.

The response from Anthropic on the Claude usage bug is that we are apparently all 'using it wrong' Thousands of people have noticed the same issue and tweeted the same token burn spike. But we're being gaslit. This is one of the worst PR I've seen from an AI company. Just blame the user and tell them to use a cheaper/shittier model.

@bollineni1234 Dude Codex on paid plan is literally INSANE. I think i got nearly 20x more usage out of it compared to claude in terms of the weekly limit

claude hit the limit. opened codex (go plan), fixed all the bugs. been using it 30 mins straight. still no limit hit. and this is just the free GO plan. genuinely curious what it feels like on a paid tier. didn't believe the hype until i actually used it. codex is genuinely good. now i'm wondering, is claude overhyped, or is openai just really bad at marketing codex? considering cancelling claude max tbh. have you started using codex?

x.com/devneeddev/sta… lol they got exposed

@vbxeric ikr

@nazranf_ typical corporate bs, you're holding it wrong 😂 so we all held it wrong for months

@topagentmike007 theyre just tryin to farm impressions and likes out of anything lol

@nazranf_ what's worse is even the consumer side has drunk the kool-aid. every reply is "here's 10 ways to reduce token usage" like it's a skill issue.

@mwfowlie I only come to Claude when I feel that Codex is struggling with a task, and it usually gets the job done. Having an extra options that takes a different approach can be quite useful.

I don’t understand why people prefer Claude. I’ve been using both it and Codex everyday and Codex is far superior.

@bollineni1234 i used codex as well and 5.3-codex model is actually insane, i would say on par with opus performance, give it a shot

@nazranf_ i am just wondering why open ai or gemini is not leveraging this moment to update their models (to compete with claude models)

experiencing one of the biggest scams in the ai industry right now. woke up -> gave 2 prompts (it was related to fixing a bug) -> hit usage limit 😲 claude max plan btw.

@mrnoplanb @warpling @AppStore not even gonna give it a download

@nazranf_ @warpling @AppStore is the app shit?

Brand new app. Claims 10 Million users in the first screenshot… Hey @AppStore…

@mrnoplanb @warpling @AppStore pretty sure any traffic that app gets from this tweet will not convert to downloads, and even if it does will probably be some hate reviews lol

@warpling @AppStore you are making the app popular and it will get 10 million users, fake it till you make it

@ipla03 @AIAdsApps makes sense, now that you point it out i realize this is exactly how my spotify subscription is set up but i just never realized

@nazranf_ @AIAdsApps You are not allowing to inform the user that you can pay on a website from inside the app, if everything happens externally it’s fine. Many apps do this: X, Spotify, Netflix, etc… It is allowed, a bit uncommon for users so I would expect conversion rates to be lower tho

Next time you open an app, look at the first screen. There's usually a button: "I already have an account." Most people scroll past it. It's one of the most profitable decisions a founder can make. Here's what it really means: 👇

@RobertFreundLaw no way he is doubling down on it 😭

Many users pointing out how impressive it is to have 10M+ users on your app at time of launch. Wow!

@burakeregar Sounds cool, what does Tenjin do?

big news: tenjin is also approved for turkish government’s mobile app support program 🚀 i am now starting with meta ads. the moment i am ROAS positive, i’ll create a turkish company and move my apps there i’ll also be back in europe next month and spend at least 3-4 months indie hacking in fethiye/turkey (spoiler alert: one of the best places in turkish riviera, just google it 😉)