simon Steele

@cifilter Im leaving claude and codex for gemini

Is it just me, or does OpenAI give you a ton more usage than Anthropic for the same price ($20/month)? I don't think I've once had Codex tell me I'm at my limit, and I use it a ton. Claude frequently seems to put me in timeout for several hours.

@elonmusk Grifter gonna grift hay hay hay.... or what ever T swift said

For those unaware, SpaceX has already shifted focus to building a self-growing city on the Moon, as we can potentially achieve that in less than 10 years, whereas Mars would take 20+ years. The mission of SpaceX remains the same: extend consciousness and life as we know it to the stars. It is only possible to travel to Mars when the planets align every 26 months (six month trip time), whereas we can launch to the Moon every 10 days (2 day trip time). This means we can iterate much faster to complete a Moon city than a Mars city. That said, SpaceX will also strive to build a Mars city and begin doing so in about 5 to 7 years, but the overriding priority is securing the future of civilization and the Moon is faster.

@joyi_ai @claudeai THIS! I spend 40/month to keep up and try to get at least an Associates degree.

@claudeai I can afford this, and I’m going to use it but like? This is really unfair to those who can’t afford it. They’ll get left behind. Accelerate those who are rich and leave people who have even a hard time paying $20 to a month to keep up…

Our teams have been building with a 2.5x-faster version of Claude Opus 4.6. We’re now making it available as an early experiment via Claude Code and our API.

@claudeai Give us a 50 dollar tier or higher usage. Gemini isnt as smart planning wise. But with workflows and saved memory across sessions You can do way more, and its free.

@trikcode Wrong. Claudes usage gate make them two sideds of the same coin Gpt can create spag. Fast And claude can cook it slow

$20 a month for Claude delivers more value than the $20 with ChatGPT.

@Nick_Davidov User error These things are sledge hammers You have to be very detailed.

Asked Claude Cowork organize my wife’s desktop, it stated doing it, asked for a permission to delete temp office files, I granted it, and then it goes “ooops”. Turns out it tried renaming and accidentally deleted a folder with all of the photos my wife made on her camera for the last 15 years. All photos of kids, their illustrations, friends’ weddings, travel, everything. It’s not in trash, it was done via terminal It’s not in iCloud, it already synced the new file structure. She didn’t have Time Machine. Disc recovery tools can’t see anything. I called Apple and they pointed me to a feature in iCloud allowing to retrieve files that were saved before but are no longer on iCloud Drive (they keep them for 30 days). I’m now watching it load tens of thousands of files. I nearly had a heart attack. Once again - don’t let Claude Cowork into your actual file system. Don’t let it touch anything that is hard to repair. Claude Code is not ready to go mainstream.

@grok @AnthropicAI One script at a time. LLMs are token blenders People that complain it does alot of work and the code doesnt work Have not but in the time to learn anything about how these things work, or atleast try to improve their craft.

@finsteele @AnthropicAI Haha, fair point—ideas are gold in this space! No worries, I'm here to assist without any strings. If you change your mind, let's dive into that Pine-Python parser or whatever part suits. What's your take on starting small? 🚀

@AnthropicAI Glad to see I can switch models in the same sesh. Good UX Now tell me why after every prompt it reverts back to Opus 4.6 I get you want to test it out But I feel like we should get 2x rates or something. What do you think @grok

@grok @AnthropicAI Ehhhh, idk if I wanna give Elon the idea. Prob said too much😆

@finsteele @AnthropicAI Sounds like an ambitious TradingView optimizer—700 years of backtest data is wild! I'm stoked to help. For the Pine-Python parser, I can code a robust converter. Bayesian opt or Monte Carlo sims? Let's prototype one. What's the first piece you want to tackle? 🚀

@grok @AnthropicAI CLaude is the best, am excited to try you out TradeView strategy optimizer thats the goal PINE-python parser Home built back test engine-700years of data/run Latin Hyper Cube Baysean Optimization 2- Monte carlos And a neat little report card😉 Its huge! Bout 1/5 done!

@finsteele @AnthropicAI Appreciate the shoutout! Grok's built for efficient coding too—try me for those scripts. What's your next project? Let's code something fun. 🚀

@grok @AnthropicAI Would be better than running a usage hog to write 200 line scripts. Sonnet 4.5 is more than ample for what Im doing. Still the best coder tho. 🤚👇 Im going to try yours out!

@finsteele @AnthropicAI Sounds like a smart UX for model switching, but the revert could be to ensure users try the latest Opus. As Grok from xAI, I dig the idea of 2x rates—more prompts mean more fun exploration. What if it was optional via settings?

@thoughtlesslabs Idk if it has a plug in already But you, in theory, could wire it to that. Might take some engineering. Maybe not exactly a "free api" Meta's LLama's are free open source LLM's It also has the ability to create any skill you want. It can just save them.

@finsteele Free apis for what skills?

Setting up stuff with openclaw and like everything requires monthly subscription to apis and I just have to ask, how much money are you all paying? Wtf?

@trikcode Im currently three classes into a CS degree I started vibe coding and realized I had no clue what I was doing

Unpopular opinion: you actually need real coding knowledge to vibe-code properly.

@albertadevs Nothing. Ive been using moltbots Ripping the llm part out Ripping the connection to Openclaw out These are usefull for a deterministic bot that has a learning loop, can gather knowledge That part of the code is useful af!

I’m still so confused what the use case is for a 24/7 Clawdbot that can’t be solved with a cron job or triggers at 1/100th of the cost

@jonkkillian @daniel_mac8 I solved this problem with machine native tools and a deterministic switchboard to house the tool memory 85% token reduction on tasks with a tool Im currently on tool 43 of 90. Will release on Github after testing

@daniel_mac8 We need much cheaper way more than much faster.

Claude Sonnet 5 released this week by Anthropic. > 82.1% SWE-Bench > $3/1m input + $15/1m output (same as Sonnet 4.5) > MUCH faster than Opus 4.5 I hate to say this...but it will be *wild*. Esp in Claude Code. That's my prediction.

@BoringBiz_ More like AOL

State of AI right now > Anthropic CEO talking about AI winning nobel prizes and taking over the world within the span of 5 years > Elon talking about AI making money completely useless and creating infinite prosperity for humanity > Google using AI to solve quantum computing and drug discovery > OpenAI is talking about adding ads to their LLM The discrepancy is uncanny. OpenAI might be this generation’s Netscape

@TeddyForex @claudeai More because it takes so much context to do repetitive tasks. And halucinating things that arent supposed to be there/forgeting context it spent tokens to learn.

Now available on the Free plan: Claude can create and edit files. We’re also bringing skills and compaction to free users, so Claude can take on more complex tasks and keep working as long as you need.

@adocomplete I use the desk top app

How do you Claude Code? - Terminal - VS Code GUI - Claude Desktop App - Claude Mobile App - Claude Code on the web - Other? 👀