Octolus retweetledi
there are a lot of benchmarks that suggest 5.6 sol is the best model in the world right now, but the most reliable way to tell is that elon is obsessed with me again
English
Octolus
1.3K posts
@heyandras @AdityaTripathiD what's the reason you chose Laravel for this project?
English
I asked @AdityaTripathiD to fully manage the cloud version of shoutrrr.com, everything, server, Coolify's conf and deployment, everything.
This is a huge level up for me as a business man. 🤘
English
@imjustnewatai Fable medium has been great, but the safeguards are way too strict. Hopefully GPT handles it better.
English
OpenAI GPT 5.6 Tomorrow. A little bit better than Fable.
English
Just shipped a small proof-of-concept: SSH command execution from Cloudflare Workers using outbound TCP sockets + ssh2.
Includes buffered JSON responses, streaming NDJSON output, and a browser demo UI.
github.com/Octolus/CloudF…
English
@heyandras I've been using Fable on "Medium" and it's honestly not that bad
English
lol, i started my first fable prompt before i went to sleep since it's back.
i run out of limits with one planning and then executing the plan prompt.
English
Octolus retweetledi
@vovudebosh Gave it a try. It was unable to find relevant competitors for our business. But with a few improvements i am sure it would work great.
English
BREAKING 🚨: THIS IS THE FIRST AI SALES TOOL THAT DOES THE WHOLE JOB. PERIOD.
NOT JUST DATA. NOT JUST ENRICHMENT. NOT JUST TEMPLATES.
RESEARCH, VERIFICATION, AND A PERSONALIZED COLD EMAIL IN UNDER 2 MINUTES.
AUTOGTM IS WHAT THE OTHERS PRETEND TO BE.
English
@heyandras Would be cool with AI Integration, allowing us to use our own keys even on the Cloud version, for rewriting, or even image generation.
English
I am already thinking of the price for the cloud version.
First thought: $10/m, unlimited everything except X posts as they are still pricey, so like 300 posts/month included.
Chat, wdyt?
Andras Bacsai@heyandras
🔥 Time to enter the next phase of my business journey. Please welcome our new cool developer: @AdityaTripathiD 💜 He is helping me for a few years here and there, but now it is official. (We are a team of 5 now 🤯) And that is not all. I asked him to make the most requested community app, and he did it in a month, so here you go: shoutrrr.com - An open source social media scheduler app. It is easy to self-host (literally one command) and backed by our philosophy of coolLabs. Currently only available for self-hosters, but soon the cloud version is coming. Lots of things needs to be added and improved, but it is in a very good shape for a MVP. I will keep you posted. Btw: this post was scheduled from Shoutrrr.
English
Great work, i started on something similar but couldn't find time to complete it.
A few ideas that I added to mine:
- AI Agent that can trigger if nobody responds, or on first, customizable rules.
- AI Agent can handover chat on certain conditions.
- AI Agent can block users from chatting based of fingerprint/IP
- Super customizable rate limiting / actions
English
@jamesqquick OpenIntercom i guess - haven’t thought of a name yet.
English
@heyandras Has the ability to revert changes like Cursor git mode been added yet?
English
🛠️ Jean 0.1.54 - 2 weeks of improvements, features, and fixes.
- Mr. Robot (beta) - auto-investigate and fix issues regularly, even set off hours, so it can do it while you sleeping.
- Command Code CLI (beta) support
- PI CLI (beta) support
- WSL support
- Steer support for Codex, Opencode, PI, so you can add new thoughts in the middle of the prompt.
- Change AI providers between prompts.
And bunch more.
English
@CristiVlad25 how you getting opus to do that, it almost everything for me
English
@AgenticLeo Yeah I noticed when i switched it lost all the conversation history, although it managed to get a good grasp of it by running a goal to analyze the past few days of commits lol
English
@OctolusNET My problem is that I have everything in wsl :( so I can't just switch
English
@xjtforu @thsottiaux I set it to "Fast" and "Medium" took over 5 minutes to change a few characters in a single file.
English
@xjtforu @thsottiaux I noticed the same, perhaps it's a method to make it feel like your subscription last longer but in theory it's taking x10 longer to even do anything 😅
English
Codex papercuts 📉
Codex adoption 📈
OpenAI Developers@OpenAIDevs
Today’s Codex quality-of-life updates start in settings. You can now search Codex settings, with results grouped by category, so you can find what you want to change without scanning every section – this makes setup and customization easier.
Français
I gave DeepSeek v4 the vulnerable code of Zcash and the same prompt, but it wasn't able to find the bug
English
English
@mkurman88 Been testing the M3 a lot, it's extremely good especially for the price. Also very good at security.
English
This is real. For the first time, I trust a model other than GPT or Claude with heavy agentic tasks in zorai. This is M3.
English