Traffy

pfp: @BroHallaNFT

[Traffy @TraffyBRO]

Shorting $Launchcoin here

it's all talk and no walk imo

[Molecule @Molecule_sci]

The White House published a report this week on the future of research funding. It calls for backing individual scientists over institutions, with open infrastructure that makes scientific contributions traceable and auditable.

Sounds familiar?

[Traffy @TraffyBRO]

@warobusiness @heart_

[Bio Protocol @BioProtocol]

June in the Biosphere 🌐

• The first wet-lab resulsts for @peptai_ are in. AI-designed binders against a cancer target, tested and already feeding the next round.

• BIOS ( @bioaidevs) can now take a protein target to ranked drug candidates in under an hour, for under $7.

• OpenLabs is live: the coordination layer where scientists post ideas, communities fund them, agents help run them, and projects graduate to the launchpad.

• $1M seed closed for ARTAN Bio, moving its suppressor tRNA platform toward first-in-human studies.

• Retroactive staking rewards went live across five projects: FOXO3, Holi, GALE, RELIEVA, and AUBRAI.

The loop is closing. Read the full report ↓

[Traffy @TraffyBRO]

@rektober Yeah agree bro. Luck is always on their side.

[Traffy @TraffyBRO]

@rektober Told ya Real does some vodoo shit. Missed out on 90k just cause of Arsenal

[Traffy @TraffyBRO]

@rektober Arsenal 4-1
City 1-2 Real

[Traffy @TraffyBRO]

@rektober Most likely yeah, hence the high odds for Real.
But Real will counter with some vodoo shii

[Traffy @TraffyBRO]

@diofarter Hello

[Traffy @TraffyBRO]

@LudwigAhgren Hiding from ICE

[Traffy @TraffyBRO]

@based16z @nikitabier @inversebrah

[BIOS @BioAIDevs]

We used a @bryan_johnson tweet to build a longevity supplement with AI agents.

Bryan posted that resting heart rate might be the single most important health marker. Fair question: what would it take to build a supplement that actually moves the needle?

So we gave BIOS - our new AI scientist - a single prompt derived from Bryan's post about lowering resting heart rate.

What the agent did:

BIOS operates in 3 modes. We ran autonomous - up to 20 iterations of literature search, analysis, and synthesis. It stopped at iteration 5 because the workflow hit a stable evidence plateau and had enough support to report back.

BIOS identified a Tier 1 formulation:

• Taurine: 1.5-3g/day (most trials used 3g)
• Omega-3 fatty acids: 2-4g/day
• CoQ10 and Magnesium as Tier 2

Each came with BPM reduction data, time-to-effect, drug interactions, and safety profiles. The kind of detail you'd want before actually testing something.

Then we ran it again. Same prompt, zero changes.

Second run: 6 iterations instead of 5. Slight variance in tier rankings (Omega-3 moved to Tier 1, Magnesium to Tier 2). But the core findings were nearly identical: same supplements, dosing ranges and mechanisms.

The convergent outputs from identical inputs - across variable iteration counts - suggest the agent is finding real signal in the literature.

After the investigation, BIOS asked: "Should I design a clinical trial for this? Calculate costs? Build a personalized protocol?"

The agent can now:
• Design wet-lab experiments with timelines and cost estimates
• Build dosing protocols based on your specific parameters
• Assess novelty potential of the formulation
• Map out contraindications and drug interactions

This is real-time biological coding: You write the spec (your research question), the agent writes the implementation (evidence-based experimental design).

All this runs on our deep research framework, which just ranked #1 on BixBench (the @FutureHouseSF data analysis benchmark).

A key feature connects the dots: persistent memory across multi-agent workflows. The literature agent, data analysis agent, and orchestration layer all share context. Each iteration builds on the last. The agent knows what it learned 3 steps ago.

[Bio Protocol @BioProtocol]

"AI bot that's doing protein folding and literally coming up with cures, doing personalized medicine for cancer patients... you could have the equivalent of like a GoFundMe, but on the blockchain, for people to basically be able to pay an AI bot to cure their cancer."

BIOS is built for this.

It generates discoveries.
Bio Launchpad funds them.
Biofy brings validated treatments to patients.

Discovery -> funding -> product

One system.

[Traffy @TraffyBRO]

@AzFlin Noted we won’t

[Traffy @TraffyBRO]

@SilvXBT @inversebrah hang it on the smolchain

[Traffy @TraffyBRO]

@MarcelRuschel @ToIgaFCB So muss es sein

[BIOS @BioAIDevs]

AI for bio has a scaling obsession.

Yesterday on stream, @SynBio1 (ex @Ginkgo, Harvard PhD) made a point that stood out while demoing BIOS, our new AI scientist.

Most of the scientific AI ecosystem is chasing a familiar playbook:
build gigantic nine-figure datasets, train huge models, predict everything.

Jake’s excitement was about something way more accessible.

He ran BIOS on a metabolic engineering prompt and it produced a yeast program for sandalwood oil:
• market pricing ($500–$2,500/kg)
• techno-economic targets
• a peroxisome scaffolding strategy
• freedom-to-operate / IP review

Then it did the important part:
it produced a wetlab experiment you can actually run.

Engineer 4 strains, run 120-hour fed-batch fermentations, measure output via mass spectrometry. Cost: tens of thousands of dollars.

Jake’s point: human scientists learn from tiny datasets all the time. Biology advances through “touching nature in small bites” and tight feedback loops.

BIOS is built for that loop.

[Traffy @TraffyBRO]

@BillyM2k @NickOzeanVII

[Traffy @TraffyBRO]

@suitman @inversebrah observe

[Traffy @TraffyBRO]

@radbro Underrated tweet