Mr.Anderson

Software Engineer

[ali(gnment) @waterloo_intern]

https://t.co/aivc7B4A7Z

[Mr.Anderson @anderson_tppnt]

Been thinking a lot about long-term memory for AI agents lately.

Everyone has a different answer: raw text, vector embeddings, or dedicated memory frameworks like Mem0.

Then MemoryArena (ICML 2026) came out and showed that all of them perform surprisingly poorly.

The reason is interesting: it evaluates how well an agent uses memory to guide future actions, not just whether it can remember.

My takeaway isn't that today's memory systems are bad. It's that we still don't know how to integrate external memory into agent workflows effectively.

This feels less like a memory problem and more like an engineering problem—retrieval, context composition, and harness design are becoming as important as the model itself.

[Arang @Arang_Persona]

https://t.co/wyA2749VOp

[codila @0xCodila]

Two Anthropic seniors just made Karpathy's loop 1000x better with "Graph Engineering" - dropped 11-page PDF

the shift: the agentic systems got 1000x better the moment you wired agents into a graph

here's the playbook in 6 steps:

step 1 → build one loop: generate, critique, revise - one self-review cycle beats a smarter model with none

step 2 → add tools: search, code execution, database - thinking without tools is hallucinating

step 3 → go parallel: spin up agents in separate worktrees - same repo, different branches, no conflicts

step 4 → add a graph: agents write findings as typed nodes and edges - not transcripts - every claim keeps its source

step 5 → ground your evaluator: it checks claims against graph edges, not vibes - "Triple not found" beats "seems off"

step 6 → the graph survives every session - your agents stop rebuilding context from scratch

the result: Karpathy ran 1 agent in 1 direction - this system runs 1,000 with shared memory - same model, it's the architecture

read this 11-page PDF and paste it into your Claude - you won't regret it

bookmark - then read the article on building graphs from scratch ↓

[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

[Arang @Arang_Persona]

🎉 Second Event Alert! 🎉
Retweet 🔄 & Like ❤️, then drop your OverFlex wallet address in the comments!

💰Create and promote your own Meme Token!

👉 How to join:
1️⃣ Use Mint Club to launch your own meme token
2️⃣ Promote it on X (Twitter) in a fun way & include these tags:
#overprotocol #mintclub #memecoin #overmeme

✨ Once done, Like & Retweet this post and comment your wallet address to complete your entry!

🎁 Prize List: OverFlex RWA tickets

Zast the Hip ticket
Over Protocol T-shirt ticket (L/M)
Convenience store coupon ticket (~$2)

[Arang @Arang_Persona]

Stake OVER,
Claim OVER,
Use OVER in real world

[Safe.eth @safe]

https://t.co/IthUlY8y3d

[LayerZero @LayerZero_Core]

Introducing Quick Merkle Database (QMDB), a high-performance verifiable database optimized for blockchains.

• 2.28M state updates per second, 1M TPS (benchmarked transfers per second).

• Benchmarked with workloads up to 15B (10x Ethereum’s 2024 state) and proven capacity to scale to 280B entries on a single server.

• Single read per state access, O(1) I/O for updates, and in-memory Merkleization on a footprint as small as 2.3 bytes per entry.

• Efficiently scales across both consumer grade and enterprise hardware.

Developed by LayerZero Labs, QMDB is MIT and apache-2 licensed and completely open-source.

[Mr.Anderson @anderson_tppnt]

Airdrop hunters keep whining about the price, but Over Protocol’s stability is getting stronger thanks to stakers worldwide. Devs are catching on and hitting us up for partnerships. Build first, price follows.

[Ben | Over 🌐 → x402 with Agentic Commerce @overnance]

A new chapter is on the horizon.
Approximately 12 million OVER tokens are palm-staked, and around 7 million tokens are home-staked. This represents about 2% of the total token supply currently staked within the protocol, with the numbers steadily growing.
Early next year, the Validator-Collateralized Stablecoin and RWA marketplace will launch, igniting a flywheel effect to accelerate the value of OVER tokens. Keep staking OVER and get ready to welcome an exciting year ahead!

[Ben | Over 🌐 → x402 with Agentic Commerce @overnance]

A new chapter is on the horizon.
Approximately 12 million OVER tokens are palm-staked, and around 7 million tokens are home-staked. This represents about 2% of the total token supply currently staked within the protocol, with the numbers steadily growing.
Early next year, the Validator-Collateralized Stablecoin and RWA marketplace will launch, igniting a flywheel effect to accelerate the value of OVER tokens. Keep staking OVER and get ready to welcome an exciting year ahead!

[Ben | Over 🌐 → x402 with Agentic Commerce @overnance]

https://t.co/tdpAe2cQbJ

[Mr.Anderson @anderson_tppnt]

One by one. $OVER is proceeding

[Mr.Anderson @anderson_tppnt]

TGE, anyway...

[Mr.Anderson @anderson_tppnt]

Blockchain is a means rather than an end.

[Ben | Over 🌐 → x402 with Agentic Commerce @overnance]

https://t.co/AVm8Ti7lNk

[Mr.Anderson @anderson_tppnt]

The delay was not by technical reasons or any bad cause. Relax and wait. The news will come out soon.

[Over @overprotocol]

Delay of the Grand Open

We are deeply thankful for the overwhelming interest and patience shown by our community🙇‍♂️

Having completed the final tests for our mainnet launch, we are now enthusiastically preparing for the grand opening!

Initially, we aimed to launch the mainnet and conduct the Token Generation Event (TGE) within June. However, discussions regarding strategic partnerships, including Sybil detection, and further collaborations have extended beyond our initial timeline, necessitating an additional delay in our launch schedule.

We sincerely apologize for this disappointing update, as we recognize the trust our community has placed in us.

Our commitment remains firm, and we are actively simulating the successful operation of our mainnet, forging partnerships, and planning exciting, profitable events that will enhance our ecosystem's growth.

We will continue to provide updates on the mainnet launch, product developments, and details about the Nethers NFT and airdrop alongside the grand opening announcement.

We appreciate the community's continued support and understanding. We are planning additional benefits to reward your patience.

We will notify you as soon as the negotiations conclude, so please wait a little longer. Stay tuned for news about various partnerships! Thank you.

[Ben | Over 🌐 → x402 with Agentic Commerce @overnance]

OverProtocol's Roadmap Unveiled @overprotocol
https://t.co/ffiQ4D183Z

[Mr.Anderson @anderson_tppnt]

@peter_szilagyi Thank you for your the interesting tweet. but I have one question: Is there any relationship between receipts and snap sync support? I thought that in snap sync, accounts are needed and that receipts are not essential.

[Over @overprotocol]

Schedule Change Notice🚨

1⃣ The @flipster_io Sybil Detection will end on Wednesday at 03:00 PM UTC, earlier than scheduled.

2⃣We are going to have a new Sybil Detection round with a new CEX in June.

3⃣So, don’t worry if you can’t participate in the current Flipster Sybil Detection.

P.S. If you are in a country supported by Flipster, please try to complete the Sybil detection process before the deadline to verify eligibility for the Airdrop.
✅ Everyone who has completed and will complete Sybil Detection through Flipster before the deadline will be eligible for the Airdrop.

Thank you for your understanding🙇

[Mr.Anderson @anderson_tppnt]

There will be news about sybil detection. We apologize for any inconvenience and thank you for your patience.