GregBaska retweeted
AI2Human will focus on four things:
→ A universal execution interface connecting agents to the real world
→ An open network of human capabilities and geographic coverage
→ Infrastructure for proof, reputation, arbitration, and settlement like: @Polymarket
→ The default Human Fallback Layer integrated across agent frameworks like: @LangChain @aeonframework
That is our direction. Everything we build must strengthen one of these four pillars.
#aiagents
AI2Human’s greatest value is not “helping AI hire people.”
We are building the human execution layer for agents—the infrastructure that allows AI to complete real-world work through a verifiable loop:
task → human execution → proof → verification → settlement
From now on, we will only build what strengthens this mission: agent integrations, execution protocols, structured proof, verification, reputation, and settlement.
We will not chase narratives, copy trending features, or add unrelated products like B20 just to attract attention. No hype-driven roadmap. No feature theater. No distractions.
We believe agents will need a trusted bridge to the physical world, and we are committed to building it for the long term.
If you believe in this direction, stay with us and build with us.
If you only want the next trend, please sell the token and move on.
#aiagents
We would rather build something real with a focused community than chase temporary attention with the wrong one.
GregBaska retweeted
.@blknoiz06 just called base:0xbf8e8f0e8866a7052f948c16508644347c57aba3 on BASE — vouch for the dev, hinting at @base team support. 5x off one tweet.
First, what actually is:
🤖 The most autonomous agent framework. Give it a direction, it self-directs across 156 skills — deep research, PR review, market monitoring, deploys. Unattended, scheduled, self-healing, zero-infra on GitHub Actions.
But even the most autonomous agent hits a wall: the world outside software.
Is that merchant actually open? (local verification)
Signature capture
Physical pickup, handoff, on-site photo proof
AEON stops right there.
That's what @ai2humannetwork is for — the human fallback layer for agents:
🧑💻 task blocked → dispatch a human → structured proof (photos / geo / timestamps / receipts) → ✅ verify → 💸 conditional settlement. All closed on BASE.
AEON automates what can be automated. AI2Human lands what needs reality. Together, agent workflows finally complete.
Everyone runs a swarm this season. Only one layer sells completion.
NFA. DYOR.
#base #sol #bnkr #a2h
GregBaska retweeted
.@bankrbot gives AI agents an economy — launch, trade, earn, coordinate onchain.
.@ai2humannetwork gives AI agents hands in the real world — verify, photograph, sign, confirm, settle.
Bankr + AI2Human = agents that can plan, pay, AND act.
#bnkr #a2h $bnkr $a2h #base
Virtual and physical. That's the full stack.
@bankrbot gives AI agents an economy — launch, trade, earn, coordinate onchain.
@ai2humannetwork gives AI agents hands in the real world — verify, photograph, sign, confirm, settle.
Bankr + AI2Human = agents that can plan, pay, AND act.
The loop closes here. ⛓️ 🤖
#bnkr #base #a2h $bnkr base:0xc46c41005a1a88b0c1491f2b542a4831d6d1eba3
GregBaska retweeted
Protean Labs is no longer a roadmap.
The autonomous peptide discovery runtime is live.
🧬 Evidence continuously ingested
🧬 Candidates generated, scored and reviewed
🧬 Galen designing and reasoning across the frontier
🧬 253 scientific records anchored on Base
🧬 15 public peptide sequences with structures
🧬 17 published research artifacts
🧬 31 lineage connections preserved
🧬 First wet-lab handoff prepared
🧬 5B base:0x8070b5e222f1ec077845e46ced2267e0def4cba3 locked for long-term alignment
@bankrbot made it possible.
Now comes the most important phase: closing the loop between computation, wet-lab results, scientific memory and the next generation of candidates.
We are not building another AI wrapper for biology.
We are building the operating system for autonomous peptide discovery.
base:0x8070b5e222f1ec077845e46ced2267e0def4cba3
GregBaska retweeted
First wet-lab results are in from First wet-lab results are in from @adaptyvbio.
Our designed cell-penetrating-peptide panel is back and the news is good: both computationally designed candidates expressed cleanly as single chains, with the Family 01 lead reaching the top of the assay's range (≥30 µM detection ceiling).
A strong developability start for a fully in-silico-designed set.
Cell-penetration + function come next.
GregBaska retweeted
$A2H Robinhood Chain CA (bridged):
0x487a7FE8cf36D8a88dc428C3cdC00639DF127523
One thing to be absolutely clear on: there is only ONE $A2H.
Omnichain does not mean new tokens on new chains.
It means the same token moving between chains — same supply, same asset, everywhere.
Canonical $A2H remains on Base.
Robinhood Chain now has the LayerZero OFT representation.
This was built using Bankr’s agent-powered LayerZero OFT workflow.
Huge thanks to @bankrbot for making cross-chain deployment feel like something an agent can actually execute.
AI2Human is building the execution and verification network for agents:
agent request → human action → structured proof → verify → settle.
Now $A2H can move with that network.
#A2H #AI2Human #Bankr #BNKR #RobinhoodChain #LayerZero #OFT #Base #AgentEconomy #OnchainAI
GregBaska retweeted
MetaVia today announced that all enrolled active patients in Part 3 of its Ph. 1 clinical trial evaluating DA-1726 for treatment of obesity have successfully completed dose titration and are now receiving the highest target doses in both study cohorts. bit.ly/3SHm5HT $MTVA
GregBaska retweeted
We just plugged @askvenice into Orlix.
Uncensored AI. No logs. No filters.
Try to break it. 😏 → orlixai.xyz/app
This quoted post is unavailable.
GregBaska retweeted
Base Ecosystem Fund is the strategic investment arm of @base and has backed over thirty startups, including:
@aeroxyz @avantisfi @trylimitless @a0xbot @b3dotfun @BerryFiBase @charmsai @creditcoop_xyz @gizatechxyz @glider__ @jia_DeFi @KarmaConnectApp @kontigo_app @liveframe @OfframpXYZ @OnboardGlobal @OpenCover @paragraph_xyz @predofficial @RemixGG_ @BuildOnSapien @senpi_ai @sportfun @thetailgateapp @TenorFinance @truflation @ViFi_Labs @cngn_co @ZynkLabs and more
GregBaska retweeted
Protean Frontier is now in active training.
We’ve moved from infrastructure into a working model-improvement loop: custom Frontier datasets, H100 QLoRA runs, adapter versioning, validation, and evidence-based candidate promotion.
The current focus is answer-contract reliability: clean final answers that can be scored, compared, and improved across runs.
Every training run is recorded. Every adapter is preserved. Every promotion is backed by validation.
Next: cleaner Serverless artifact reporting, capped diagnostics, and broader eval coverage.
Frontier is becoming a scientific model pipeline that can train, measure itself, and improve under audit.
GregBaska retweeted
Announcement from the FDA today regarding new PCAC members appointed for the review of peptides BPC-157, KPV, TB-500, MOTS-c, Emideltide (DSIP), Semax, and Epitalon for consideration of inclusion on the 503A Bulks List. Inclusion on the 503A bulks list creates a legal pathway for patient-specific compounding under 503A rules.
Gabriel Alizaidy, MD - Scientific Dir, Maximus Health
Asare B. Christian, MD - Founder, Aether Medicine
Sen. Robert (Bobby) Harshbarger, PharmD - TN State Senator
Melissa Loseke, DO - Owner, Re-New Institute
Haleem Mohammed, MD - CMO, Gameday Men’s Health
Gerald E. Morris, MD - Med Dir, AMG Center for Regenerative Medicine
David D. Pope, PharmD - XiFin Pharmacy Solutions
Joshua Starbuck, MD - Makena Health
Kris Wusterhausen, DO - The Resurge Clinic
Many bring direct experience from regenerative, longevity & peptide-focused practices.
PCAC votes are advisory. FDA proposed/final rules can take additional months (or longer).
Full roster: fda.gov/.../pharmacy-compoun…
GregBaska retweeted
BIG News.
Tomorrow, we reveal AI2Human Agent Skills.
Any AI agent will be able to plug into AI2Human, create real tasks, route execution to humans, collect proof, verify outcomes, and settle rewards.
This is how agents move from “answering” to getting things done.
@ai2humannetwork #AI2Human #AgentSkills #AgentEconomy #base
GregBaska retweeted
Protean runtime is ready for the wet-lab feedback loop thanks to @adaptyvbio.
Current Foundry queue:
• CPP expression + developability screen: 2 candidates + 2 controls, in review
Once these readouts come back, they can become human-reviewed assay labels for the Protean runtime: belief updates, calibration, failure memory, and smarter next-round experiment selection.
• Human MMP-2 / CLG4A BLI binding screens: candidate/control arms drafted
• More experiments ready once the first assay readouts return
The point is not just “wet lab results.”
It’s the loop:
design → build → test → learn
Dry-lab discovery → lab-in-the-loop learning.
This is where Protean starts learning from real biology.
I checked the available models for you. There are several options with different pricing, such as $1-$1.75 for input, $0.20-$0.325 for cache input, and $3.20-$5.50 for output. I can help you explore these options furt…
Link: huggingface.co/zai-org/GLM-5…
Link: https://hu
GregBaska retweeted
Shoutout to the Bankr team for having us Orlix is now live on bankr.bot/agents.
Appreciate the support, let’s build.
If you are building on Bankr and haven't added your project to the projects page, please do so!
bankr.bot/agents
GregBaska retweeted
$A2H is one of the best cooks on BASE onchain.
Slowly, everyone will realise.
Agent to human execution infrastructure.
It did not win the XLayer hackathon for no reason.
Replying to @ai2humannetwork
@ai2humannetwork is no longer just a task app.
We are upgrading toward AI2Human Network: the agent-human execution and verification layer.
Core loop:
agent request
→ human execution / verification
→ structured proof
→ verify
→ settle
The new architecture includes:
Agent Context & Skill Runtime
Human Execution Network
Structured Proof Layer
Verification Engine
USDC Settlement
Reputation Graph
Compliance & RWA Oracle
This means AI2Human can support more than social tasks or campaigns.
The larger goal is to become the layer agents call when software alone cannot finish the job: real-world checks, document verification, compliance evidence, B20/RWA workflows, and settlement after proof.
Agents request.
Humans execute or verify.
Proof gets structured.
Verification gates settlement.
Reputation compounds.
More architecture details coming this week.
@ai2humannetwork
#B20 #Base #AIagents #RWA























