Aditya Vasireddy

I build AI agents and kill them when they miss their numbers.

[Aditya Vasireddy @KillCriteria]

Every cleared context creates a choice.
Either the next session starts from understanding…
or it starts from assumptions.
A strong session handoff doesn't document every change—it preserves the reasoning behind those changes. See comments for the full details!! #AI #claudeSkill

[Aditya Vasireddy @KillCriteria]

https://t.co/r8bA8sRkV1

[Aditya Vasireddy @KillCriteria]

Every cleared context creates a choice.
Either the next session starts from understanding…
or it starts from assumptions.
A strong session handoff doesn't document every change—it preserves the reasoning behind those changes. See comments for the full details!! #AI #claudeSkill

[Aditya Vasireddy @KillCriteria]

Explore the skill:
https://t.co/xuFXGyDrxL

[Aditya Vasireddy @KillCriteria]

@Diego_exits edit

[Aditya Vasireddy @KillCriteria]

Claude plans. Codex executes. Claude verifies.

A guide to connect Codex to Claude Code through MCP, keep skills separate, and build a reviewable workflow:
https://t.co/1SarnlFoFS #ClaudeCode #Codex #MCP #aiengineering #AI

[Aditya Vasireddy @KillCriteria]

VIBE CODED: The Five-Gate Method: A Practical Framework for Bu... https://t.co/avGARaz99g

[Aditya Vasireddy @KillCriteria]

Explore the framework:
https://t.co/4h8HBrARRJ

[Aditya Vasireddy @KillCriteria]

The Five-Gate Method adds a simple discipline for building more reliable AI agents:
✅ Scope the task
✅ Inspect real evidence
✅ Challenge the first answer
✅ Verify the outcome
✅ Report uncertainty honestly
Reliable AI isn't about bigger models—it's about better execution. #AI

[Aditya Vasireddy @KillCriteria]

Claude’s automatic context compaction helps long-running AI agents avoid context overload by summarizing history, preserving key progress, and continuing tool-heavy workflows efficiently. Guide:
https://t.co/YTQOdF94EF

[Aditya Vasireddy @KillCriteria]

https://t.co/kL8kDkbhvG

[Aditya Vasireddy @KillCriteria]

Most developers use Claude Code the wrong way.
Don’t give it your entire codebase and say:
“Fix everything.”
Use these 5 rules instead:
✓ One task at a time
✓ Keep CLAUDE.md short
✓ Ask it to explore first
✓ Run tools and tests
✓ Review every change
#ClaudeCode #AI #Coding

[Jensen Huang @JensenHuang]

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.

AI will transform every industry, power every company, and be built by every country.

Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.

The world needs both frontier closed models and frontier open models.

https://t.co/AUKzoQ5Ikb

[unicode @unicodef1wn]

A senior Anthropic engineer leaked his internal PDF on giving multi-agent systems permanent memory.

Your agent forgets everything the second its context window closes. His fix: a knowledge graph that survives past any single run.

Extract → Resolve → Assemble → Query → Repeat:

Extract: Haiku pulls entities and triples, one call per doc
Resolve: Sonnet matches "Edwin Aldrin" to "Buzz Aldrin" with zero string overlap, just context
Assemble: canonical nodes, typed edges, provenance on every triple
Query: Sonnet reasons over the graph, every answer points to a specific edge

Wire it in as shared memory: workers write to it, evaluators check claims against it, loops run overnight without losing state.

Save this before it gets taken down 👇

[Nav Toor @heynavtoor]

https://t.co/VloDpAi6vy

[Aditya Vasireddy @KillCriteria]

VIBE CODED: I Built an AI Reviewer That Refuses to Tell Me If ... https://t.co/eA00tTiqm6

[Aditya Vasireddy @KillCriteria]

I ran my own startup idea through an AI review protocol I built. It didn't survive the review.

Wasn't a failure — that was the point.

Crucible restricts what an LLM is allowed to judge, instead of asking it for opinions. MIT licensed.

https://t.co/nyXanH4FV8

[Aditya Vasireddy @KillCriteria]

@AndrewBolis YouTube

[Aditya Vasireddy @KillCriteria]

The term "RAG" will be historical by 2030. MCP will be the HTTP of agents. Context engineering will be a staffed role. Here's the full AI stack lifecycle mapped out →

[Aditya Vasireddy @KillCriteria]

Every AI coding session I run has the same flaw.
It doesn't forget syntax. It forgets decisions.
So I built a memory layer for it �� (4 slides, repo link in reply) #buildinpublic #AIagents #skills