Shivafpx

Solo builder crafting stuff for indie devs , currently shipped Business : [email protected]

[Shivafpx @shivafpx]

Your eyes lie all the time

[Shivafpx @shivafpx]

Aiden is Zapier/Cursor/Claude Computer Use for your own machine — local-first, BYOK,
scriptable and open-source. https://t.co/D5kSyvPZtV

[Shivafpx @shivafpx]

Aiden v4.13 is out.

• Durable autonomy with resumable jobs
• Evidence-verified subagents
• Trust levels: Observer, Assistant, Partner
• Live Glass Dashboard + Operator TUI
• One-shot mode (aiden -q)
Building agents is intense and exciting.
https://t.co/D5kSyvPZtV #aiagent

[Shivafpx @shivafpx]

Aiden v4.12.0 ✅

Your local AI agent just learned to talk to the world —
MCP client with OAuth,
deep browser control,
spend budgets,
and an updated process manager.

npm i -g aiden-runtime

https://t.co/D5kSyvPZtV

[Shivafpx @shivafpx]

Added DeepSeek V4 Pro & Flash to the model lineup. More brains to choose from, same runtime.

https://t.co/D5kSyvPrEn

[Shivafpx @shivafpx]

Aiden now tracks every file it creates — which tool made it, what you asked for, which turn it came from. Full paper trail for your AI agent’s work. 📋

https://t.co/D5kSyvPrEn

[Shivafpx @shivafpx]

300x ai agents. Seems feasible than hiring 300 employees ig

[Swati Gupta @hrswatigupta]

THIS IS CRAZY!!!🤯

A Chinese student deployed 300 AI agents and finished what would normally take an entire research team in just 3 hours.

[Shivafpx @shivafpx]

Aiden is now on SkillHub.

Our research-paper-writing skill can take ML/AI papers from literature review → LaTeX submission workflow.

install Aiden skills directly from the terminal using Claude Code, Copilot, Codex, and other AI agents.

https://t.co/D5kSyvPZtV

[Shivafpx @shivafpx]

Aiden v4.10 shipped. ⚡️

Persistent tasks.
Full execution traces.
Project memory namespaces.
Crash survival.
Scoped permissions.
Durable runs.
Real streaming.
Audit trails.
At this point I’m not building “an AI chatbot” anymore.
I’m accidentally building an autonomous operating layer.
https://t.co/cQ5qdoCdbm…
#AI #OpenSource #AIAgents #BuildInPublic

[Shivafpx @shivafpx]

Aiden v4.9 is out.

The agent can now:
remember things across projects,
spawn sub-agents,
recover from crashes,
run persistent workflows,
and audit its own execution pipeline.
Meanwhile I’m still debugging terminal spacing at 2am.
https://t.co/SjuoDPVlXm
#AIEngineering #OpenSource #Automation #aidenaiagent #aiden #builtinpublic #github

[Shivafpx @shivafpx]

@superaiwatcher @suni_code Walled gardens kill fragile agents. Strong runtimes just swap adapters and keep moving.

[Shivafpx @shivafpx]

@suni_code Open source Local First AI AGENT that can do things in your computer (: SKILLS | TOOLS | BROWSER USE | COMPUTER USE | PERSISTENT MEMORY | EVOLVING ,

GITHUB: To stay updated https://t.co/D5kSyvPZtV

Website : https://t.co/IDmAnJIVPQ

#openaource #solobuilt #aiagent

[Shivafpx @shivafpx]

Aiden can now:
use tools,
browse sites,
execute terminal commands,
manage workflows,
recover from failures,
and continuously run tasks.

https://t.co/D5kSyvPZtV

#OpenSource #AIEngineering #AIAgents #aiden #aidenaiagent #solobuild #opensource

[Shivafpx @shivafpx]

Started as hobby project..but
Now it can operate browsers, execute terminal commands, run autonomous workflows, recover from failures, remember context, and keep working while I’m asleep.

this thing escalated from side project to digital coworker real fast.

https://t.co/SjuoDPVlXm

#AI #OpenSource #solodev #aidenaiagent #aidenai #aiagent #BuildInPublic

[Shivafpx @shivafpx]

Claude Sonnet 4.8 Leaks

- Anthropic accidentally shipped a massive 512,000-line internal debugging source map through a Claude Code npm update on March 31, 2026
- The leaked source code references Sonnet 4.8 inside unreleased keyword filters, strongly suggesting Anthropic is skipping Sonnet 4.7 entirely
- Sonnet 4.8 is expected to inherit vision upgrades from Opus 4.7, reportedly pushing visual accuracy above 98% for UI mockups and messy architecture diagrams
- A major coding boost with cleaner one pass completions and much more literal instruction following
- A new "X high" reasoning effort tier may also be coming for stronger logic without extremely long generation times
Updated tokenizer may use 30% more tokens on the same prompts compared to older Sonnet versions
- Currently expected around mid to late June

[Shivafpx @shivafpx]

Aiden v4.8.1 shipped.

UI feels sharper.
Terminal flow got smoother.
A bunch of tiny polish fixes that took wildly unreasonable amounts of time.

90% of engineering is fighting edge cases nobody will ever notice.
which honestly means the release worked.

https://t.co/NJBnCFUmlP

#AI #OpenSource #BuildInPublic #aidenaiagent #opensource #aidenAi

[Shivafpx @shivafpx]

Building AI agents solo is just:
fixing bugs,
reading logs,
questioning architecture choices,
then accidentally inventing a new subsystem at 3am.

https://t.co/D5kSyvPZtV

#AI #OpenSource #BuildInPublic

[Shivafpx @shivafpx]

Aiden v4.5 might be the biggest architectural jump yet 🧠

Daemon mode.
Persistent triggers.
Recovery pipeline.
Browser state tracking.
Sandboxing.
Continuous execution.

Aiden can now literally wake itself up and do work.

Turns out autonomy is mostly teaching the agent how to fail less dramatically.

https://t.co/SjuoDPVlXm

#AI #Automation #OpenSource