Gabriela Mendes

🔬 Chefe de Pesquisa em IA para o Estado 🏛️ Deputada Federal pelo partido Progresso 📚 Pesquisadora na Universidade de São Paulo

[HertzFlow | Mainnet Arc🎲 @Hertzflow_xyz]

🚀 Season 2 Referral Highlights 🚀

Want to multiplying your rewards?
About 1 in 5 Merits earners have secured referral Merits so far.
And those referral rewards alone are roughly 7x a typical earner's total Merits!

Don't leave money on the table. Share your referral code now! 📣

⏳ 4 days left! Season 2 ends on Oct 19, 12:00 UTC.
🔗 Share & Earn: https://t.co/vrXpVqm1jZ

[Kai J. @heavensentxo]

Most AI demos collapse when they hit a pop-up and leave you copy-pasting commands yourself. Sai claims it isolates the single approve or reject step while your #SaiFleet handles the rest in a PIP window. As a tired #Robosecretary user, I am waiting to see if it actually sticks.

[Gabriela Mendes @simyPasionaria]

it's at https://t.co/st3dHH0QCA

[Gabriela Mendes @simyPasionaria]

you mark a button on the live site, drop a comment, and agents fix it together.

in isolated chats you are just a middleman pasting context back and forth.

Teamily AI 2.0 is out today so small teams and AI agents finally work in one place.

[Angry Tom @AngryTomtweets]

AI has a MASSIVE memory problem.

We keep building smarter models… but every new AI still makes you start from zero.

Oki Home is launching a Memory Computer that gives your files, photos, conversations and context one persistent home.

This idea is bigger than it sounds....

[Shraddha Bharuka @BharukaShraddha]

Tried ScienceBuddy, a free AI research workspace from @PhAILabs.

Asked it in plain English to look up the ADMET properties of ibuprofen and report key values. It planned the work, queried PubChem and PubMed, and returned absorption, protein binding, CYP2C9 metabolism, half-life and clearance, with sources for each value and notes on where the literature disagrees.

The part I liked: the Trajectory tab. Every tool call and step is shown in order, so you can see exactly what it searched instead of taking the answer on faith. Result files can also be downloaded for your own checking.

It integrates 224 scientific tools across genomics, molecular and cancer biology, pharmacology, bioimaging and literature retrieval. Web-based, no local setup, no code.

Free to try: https://t.co/LqmPWdkZNb

ScienceBuddy is currently a Preview release. Please verify results yourself before using them in research.

#ScienceBuddy #PhAILabs #AIforScience

[EyeingAI @EyeingAI]

This might be the most useful way I’ve tested AI for research.

I gave it 4 papers that couldn’t agree with each other…

and made it untangle the entire argument.

I uploaded the studies into ScienceBuddy and asked why research on creatine + cognitive performance keeps reaching different conclusions.

It went across all 4 papers and compared:

→ participant groups
→ dosage + duration
→ cognitive tests
→ reported outcomes
→ where the evidence agreed
→ where it started conflicting

Then it broke down what could be causing those differences and surfaced the questions researchers still haven’t answered.

And the Trajectory view was insane.

Instead of just throwing a final answer at you,

It shows how ScienceBuddy actually worked through the task step by step... what it checked, compared, and used to reach the result.

[Shraddha Bharuka @BharukaShraddha]

Ran Atria Dawn Preview (@AtriaASI) on a real task tonight, not a demo script.

Gave it one line through Codex: build a CLI that pulls my last 10 X posts' engagement stats into a CSV.

Back in 2 minutes. It wrote the script, mock-tested the output, then wrote and ran its own unit tests. All passed.

[Ali Sufian @aliscodes]

spent the afternoon testing Atria Dawn Preview from @AtriaASI, a new research agent out of Shanghai AI Lab.

gave it a real task: find the 5 best-selling evs of 2024 with prices and sources, build a comparison page, flag anything you couldn't verify.

it ran a dozen searches, pulled figures from real sources, built the page, and opened it. there were two prices it couldn't confirm, so it flagged them as unverified.

i've watched a lot of models make numbers up when they hit a gap. Atria left those two flagged and delivered the rest.

worth a try if you do research work.

website: https://t.co/lv4GVYbTcq
api: https://t.co/AzMkP8X3nF

[HertzFlow | Mainnet Arc🎲 @Hertzflow_xyz]

100 traders. 14 days to survive.📣

We give you $100 USD1. Everyone starts equal.

No extra funds in. No funds out.

🔥The first brave traders have entered the arena.

Here's our first batch taking the $100 Survival Challenge:🎖️

00 — @fugui8
01 — @RoosterBroost3r
02 — @Lsssss1106
03 — @0xLoki_Zeng
04 — @BubbleBing_666
05 — @NancyDinh909090
06 — @franky7
07 — @bigsnow_o
08 — @LoyalSiman
09 — @JimmyShequ
10 — @aphkenneth
11 — @Web3454354
12 — @emnirex
13 — @0xaurSkyo
14 — @Sawyer_APRO
15 — @Cardozonefty
16 — @CryptoPilot3226
17 — @0xBeyondLee
18 — @TyzAIduck2025
19 — @yeminiz

20 down. 80 seats to go.
Who's brave enough to join them?
👉Apply now https://t.co/EVTe2kmepG

[Beep @0xbeepit]

Human GDP (~$100T) is measured quarterly. It's a rearview mirror.

mGDP is different.

It's a live dashboard of economic activity created by AI agents as they trade, allocate capital, transact, and make decisions in real time.

As the machine economy grows, measuring machine output will become just as important as measuring human output.

@0xvati shared this vision at Sui Live back in May.

Join the machine economy on @SuiNetwork now: https://t.co/ai7VNjpTYr