TheCharlieMachine retweetledi
Chris Camillo & Amit Kukreja on How to use AI for Research + Agentic Trading x.com/i/broadcasts/1…
English
TheCharlieMachine
207 posts
TheCharlieMachine
@CharlieMachineX
Building - Kin-Guard: Next-gen phone scam protection 🛡️ - XpodDaily: Your X followed feed to cool daily podcasts 🔊
Orléans - France Katılım Ocak 2022
63 Takip Edilen14 Takipçiler
@robj3d3 @tibo_maker Just a quick work on svg generation from opus to fable proves your point. 😁
English
@tibo_maker Everyone should be orchestrating with Fable 5 xhigh and implementing with Fable 5 medium
It is absolutely unfreakinbelievable
English
@tibo_maker Got both and I never understood the hype about codex. Seemed to be just average on a frontend stack, and just seemed to be okay to consume tokens without the user noticing.
But it might be a case of "it all depends on your usecase". 🤔
English
Well, hope you enjoyed #claudefable while it lasted 🫡 byebye fable 5
English
@benja_maker Already on it. Building speed has become really crazy with AI. Always trying to push the limits
English
@CharlieMachineX time for changing and bringing anything in head to life
English
this's really impressive
i don't know how accurate the copy would be but grateful it exists
Figma@figma
Imagine a world where you could copy/paste websites into editable Figma layers (jk you don’t have to imagine you can do this now with our Chrome extension)
English
A lot of people have been asking me how I marketed my app to make $200k in revenue in 3 months. If I went live for an hour to answer all of your questions would that be helpful?
English
Good times!
working on 2 projects at the time:
➡️BEMYPOD - your feed to a cool daily podcast 🔊
➡️KIN-GUARD - A very smart anti-phone scam tool ☎️
English
TheCharlieMachine retweetledi
@tibo_maker Now just imagine what will be the costs with mythos 5...
AI getting more expensive one model after the other. Would be have to have a proper AI orchestration where Haiku is would be able to spawn Fable only when needed.
English
yesterday I burned $500 in 1 prompt
fable-5 is the first model where I stop and think before running prompts, considering the cost
Bhanu Teja P@pbteja1998
It's going to be pretty sad once Claude Code subsidy goes away... There is no way I will be able to spend thousands of dollars on tokens every month... Until that happens, going to use it to the max...
English
@robj3d3 Just Opus getting closer to Haiku as an incentive for you to pay for the biggest one 😉
English
@Gavel_on_X got catapulted to 100% usage reaaaally fast 🤣
WEIRD thing yesterday 11pm france time, I was at 80% usage with a reset scheduled 4hours later. Then it went back to 0% usage... which made me work late to use it well.
Strange but thanks #anthropic !
English
The #elevenlabs voice generation is really above #Google and #openai.
Their flash 2.5 model is B+, but their V3 is A (not A+ yet). You can really set the pace, tone, and more. I think that there is potential to get a very natural result in terms of podcast-like voice.
English
UPDATE:
✅Sync with X accounts
✅Generate a podcast text from a cached X feed
✅Add the generated episodes in a list
NEXT:
➡️Test voice generation models
➡️Create a few personas with their own personalities
➡️Iron out the generated text kinks
Any input? let me know 😉👍
English
@CharlieMachineX @cursor_ai The MCP setup gets sharper when the platform itself has one. Zerops' MCP server gives Claude/Cursor full deploy and DB access, so the agent stops gluing two providers together for backend + DB.
English
What do I use for my apps right now? Tools that are all compatible with @cursor_ai :
- Database: Supabase
- Domains: Cloudflared
- Hosting: Vercel
- API: Railway
- Repo: Gitlab
- Project management: Notion
ALL connected with MCPs, looks like an orchestra. Any addition to make?
English
I've made this website for my AI training services for french SMBs. To be clear, It's a start with still quite a lot of rough edges. 😆
Tell me what you think on the UI/UX😉 (domain name is temporary)
yourlittlefactory.com
English