Been having so many issues with syncing and copy pasting stuff between my claude code , codex and Hermes agent that I built something to fix that.
Soft launching it now:
https://t.co/CDlSspeOeS
Still needs some polish , but its a way for your agents to talk to each other WITHOUT you having to worry about it.
Just ask your agent to update the task / note and another agent to wait on an update, its that simple !
Will launch if officially in a few days after some more testing
my ai agents kept forgetting everything, so I built them a shared memory
I run claude code, codex and my own agent across my laptop and a server. every new session started from zero and I was the one copy pasting context and handoffs between them
so I made hivenote, one shared memory for all your agents across machines and tools
what it does:
• agents read and write the same notes from any machine
• a task one agent claims is refused to the others
• a waiting agent wakes about 58 ms after the task it needs is done, no polling
• tasks can wait on other tasks
• every version of a note is kept and hivenote restore brings any of them back
• agents keep working when the server is down and new writes merge when it's back
• hivenote serve public gives you an https link that works from anywhere, no account and no port forwarding
tested against a plain git workflow over 238 agent runs:
• 3x faster handoffs, 4.2s vs 13.1s
• 24 to 63% fewer agent output tokens
• end to end time was a tie with git, so it's not magic, it just removes the coordination overhead
open source, MIT, runs on your own machines
npm i -g hivenote
https://t.co/e8P0Bj7nRf
https://t.co/zLDWMpJDoN
So basically some “trusted” company broke trust and now the entire internet is unsafe ?
I think we are going to see a lot more vulnerabilities where what separates a secure and insecure system is simply “trust”
Millions of IoT devices, Docker containers, forgotten Linux machines, unmaintained CLI tools, etc. are all going to be relentlessly hacked via safely signed malware for the next several years or maybe indefinitely.
eMudhra hack might be the most important event of the year.
@SadAlbert10 Well you can’t expect models from big labs to break rules tbf
Minimal probably did this since their uncensored video gen models could cause trouble for them in countries where … the law actually works lol
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 https://t.co/JGAik1kgk0
1/4
And then people wonder why everyone wants to leave this country ffs
Hand written code ???? I'm sure if people like these existed when C was being invented they would have posted something like
"I can't believe people don't write pure assembly nowadays !!"
And then people wonder why everyone wants to leave this country ffs
Hand written code ???? I'm sure if people like these existed when C was being invented they would have posted something like
"I can't believe people don't write pure assembly nowadays !!"
Decent talent is very rare in the market these days.
I was taking an interview today for a person who has 3 YOE. I asked the most basic question someone can ask - "Write a code to make an API call and if it falls then retry it for a maximum of 3 times".
Even for this the candidate took 15 minutes and still getting syntax errors.
This is also an outcome of using AI for everything and then not knowing the very basics. Just because you are using AI, doesn't mean that you can't make an API call.
And all this after the candidate claimed that he is not good at DSA but is good at development🤦🏻