Today I start three free days to travel and focus on the final steps for Vesta.
To do list before launch:
- add welcome mails
- privacy policy/GDPR footer
- add a minimum onboarding
- add some analytics to check what users use more, and what they never touch
Any suggestions for the list? 👀
@claudeai The beast has been unleashed, with clipped wings. Can't wait to find some time to test its potential in the final days before launching Vesta.
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
Thinking about the launch in 1 week and feel fear and excitement.
Fear because probably no one will care about my product.
Excitement because it's the first time I launch a product. First time I got an idea and actually focused enough to build it.
Being positive, in the worst scenario I'm trying things to improve my life. Making something is better than waiting until something magical improves my life. Most people never ship anything. I'm about to.
I don't build side projects for the paycheck.
I build them so if my job vanishes tomorrow, I'm not standing in the rubble with empty hands.
Vesta is my safety net woven in code.
#buildinpublic
Maybe we don't need Mythos to find vulnerabilities on Spanish systems... Many of the government sites are obsolete, with "inactive buttons" that you can edit on webview and send the "wrong" form without previous checks...
𝗜'𝗺 𝗻𝗼𝘁 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗮 𝗻𝗮𝘁𝗶𝘃𝗲 𝗮𝗽𝗽 𝗳𝗼𝗿 𝗩𝗲𝘀𝘁𝗮.
App Store review took days. A rejected update means a broken product.
With a PWA I ship when I want. Fix when I need to.
Solo dev. No time for gatekeepers.
#buildinpublic
Every budgeting app wants your bank credentials.
Vesta doesn't connect to your bank. Never will.
Built for couples and families who just want to manage money together — without handing over their financial life to a third party.
Manual entry is the feature, not the limitation.
#buildinpublic
Idk about Opus 4.6 because I started with Opus 4.7 and rarely use it, but since their agreement with xAI to use Colossus, token burn has decreased a lot. Before the agreement, any big prompt would consume 20–30% of the context window. Now, even with complex tasks, I rarely hit 100%.