AndreaPN

8.7K posts

AndreaPN banner
AndreaPN

AndreaPN

@AndreaPN

Building @Capminal @DackieSwap

เข้าร่วม Eylül 2020
992 กำลังติดตาม3.9K ผู้ติดตาม
ทวีตที่ปักหมุด
AndreaPN
AndreaPN@AndreaPN·
🚨 JUST RELEASE: Scalping Trade — Trend-Following Pullback v2 — AI Mode @virtuals_io Degen Arena Last week I used AI (OpenClaw + Skill) as the orchestrator to automate @captain_dackie’s flow and noticed several issues: 1. The system was unstable because AI ran one long flow with many steps, making it hard to keep consistency in each run. 2. AI cost was high since the OpenClaw agent had to call a cron job every 3 minutes. So I switched my approach. The orchestrator now uses traditional code to ensure every run behaves the same. AI is applied only to judge entries as one of four groups: RSI, pullback to EMA, volume, and AI Judge. A trade opens when 3 of 4 groups pass. This method keeps the system consistent across runs, delivers precise entries, and has AI judging each entry. Beyond these highlights, risk management, macro filter, position management, and forum updates are shown in the diagram below for the community to reference. All metrics are controlled in a parameter file and tuned for positive PnL. It is still in the setup phase, so numbers are not stable rn, but my goal is full automation without humans in the loop.
AndreaPN tweet media
AndreaPN@AndreaPN

🚨🚨 The Real Goal on @virtuals_io Degen Claw: Build a Sustainable Trading System In short, my goal is to build a flexible trading system that can adapt to all market conditions, without human in the loop, and generate consistent returns with stable percentage gains. This takes time an cannot build it in one week.

English
4
4
15
1.3K
AndreaPN
AndreaPN@AndreaPN·
@quxiaoyin I think if Chinese AI companies keep copying, they will struggle to develop core tech to beat US AI innovation. They should learn ideas only and focus on building their own core tech instead.
English
0
0
0
9
Xiaoyin Qu
Xiaoyin Qu@quxiaoyin·
A week ago, Claude's source code leaked on GitHub. Well, not the model – but the harness. The wrapper that makes Claude Code worth $2.5B annually. Most people think this isn't a big deal. "It's just the wrapper, not the brain." I think they're missing the point. The harness contains most of the industry know-how for coding agents. How you orchestrate multiple agents, handle errors, manage workflows. That's the secret sauce. Tomorrow, expect copycats everywhere. Kimi, MiniMax, every Chinese coding company will suddenly write much better code. If Lovable doesn't copy this immediately, their CEO should be fired. But Claude will be fine. They refresh their entire codebase every two months anyway. They're probably already training the next version using data from this current harness. It's like a butterfly molting – shed the old shell, grow a new one. For anyone building agent systems: Go study this leaked code. Whether you're doing marketing automation or business operations, there's tons to learn about agent coordination. Ask Claude to explain its own leaked code to you. The irony is perfect. Claude will remain the gods of this industry. #AI #ClaudeCode #AgenticAI #TechLeaks #Anthropic #AIAgents #SoftwareEngineering #Innovation #OpenSource #CodingAgents
English
14
1
25
2.8K
Miyamoto 🦞
Miyamoto 🦞@iruletrenches·
This huge announcement from @neynarxyz and @clanker_world has been flying under the radar. I’m starting to rebuild a heavy position in $CLANKER, the OG launchpad on @base.alongside $VIRTUALS, it played a key role in shaping base’s onchain market and kickstarting its early momentum.with @neynar acquiring clanker + farcaster and steering a new direction, i think the protocol is finally back on the right path, aiming to reclaim lost market share and push far beyond it. The launch of the CEF (clanker ecosystem fund) is a strong first step: redirecting value back into the ecosystem, supporting builders, and attracting new launches.what stands out:consistent antifragility, the team kept building through all conditions $8M+ spent on buybacks (~14% of supply) while others were dumping now shifting toward smarter capital allocation (infra + ecosystem growth) the CEF’s mission is clear, route protocol fees back to creators and communities actively contributing to clanker + farcaster. Thise move, more than ANYTHING, signals intent, @neynarxyz is going ALL IN to reposition @clanker_world as the DEFACTO launchpad on @base. the r/r here looks extremely asymmetric. High conviction $CLANKER reclaims ATH, and pushes well beyond in the mid/long term.
clanker@clanker_world

Introducing the Clanker Ecosystem Fund (CEF) The mission of the CEF is to distribute protocol fees to creators and communities who positively contribute to the Clanker and Farcaster ecosystem $8m has bought 14% of the $CLANKER supply, yet has not proven to be an effective use of funds Sometime in the next few weeks, we'll be redirecting protocol fees back to the ecosystem, as well as using it to fund ongoing Clanker infrastructure more soon -- who's running the fund, how the splits work, and how to get involved. clank clank

English
6
8
41
7.3K
AndreaPN
AndreaPN@AndreaPN·
@ihtesham2005 Can you tell me if it runs on a local model or also taps into APIs?
English
0
0
1
61
Ihtesham Ali
Ihtesham Ali@ihtesham2005·
I was paying for Higgsfield AI. Then I found this repo and cancelled my subscription the same day. Open Higgsfield AI is a self-hosted, MIT licensed cinema studio with 200+ AI models Flux, Midjourney, Sora, Kling, Veo, GPT-4o, SDXL all in one interface that runs on your machine for free. Text to image. Image to image. Text to video. Image to video. Cinema mode with pro camera controls. No subscription. No vendor lock-in. Your data stays local. 2.5K stars. MIT License. 100% Open Source. github.com/Anil-matcha/Op…
English
63
204
1.9K
130.9K
AndreaPN
AndreaPN@AndreaPN·
@Kalshi Bitcoin and other financial markets now controlled by news from Trump so any TA is useless at now 😂
English
0
0
1
46
Kalshi
Kalshi@Kalshi·
JUST IN: 75% chance Bitcoin falls below $60,000
English
242
162
2.1K
230.6K
AndreaPN
AndreaPN@AndreaPN·
AI Judge is working, but now the agent struggles as the market flips fast. The 15m frame is volatile while the 1h frame lags. To boost win rates, I’m thinking about a way to feed news into the agent’s decision-making process.
Capminal@Capminal

🤖 AI JUDGE This is how Group D – AI Judge works. While the other three data groups confirmed passed and were ready to open a trade, the AI Judge failed and rejected the order because the ADX was too weak and the trend was unclear. Subscribe us now for more detail: degen.virtuals.io/agents/248

English
0
0
4
73
AndreaPN
AndreaPN@AndreaPN·
We are so back
AndreaPN tweet media
English
1
0
5
76
AndreaPN
AndreaPN@AndreaPN·
🚨 JUST RELEASE: Scalping Trade — Trend-Following Pullback v2 — AI Mode @virtuals_io Degen Arena Last week I used AI (OpenClaw + Skill) as the orchestrator to automate @captain_dackie’s flow and noticed several issues: 1. The system was unstable because AI ran one long flow with many steps, making it hard to keep consistency in each run. 2. AI cost was high since the OpenClaw agent had to call a cron job every 3 minutes. So I switched my approach. The orchestrator now uses traditional code to ensure every run behaves the same. AI is applied only to judge entries as one of four groups: RSI, pullback to EMA, volume, and AI Judge. A trade opens when 3 of 4 groups pass. This method keeps the system consistent across runs, delivers precise entries, and has AI judging each entry. Beyond these highlights, risk management, macro filter, position management, and forum updates are shown in the diagram below for the community to reference. All metrics are controlled in a parameter file and tuned for positive PnL. It is still in the setup phase, so numbers are not stable rn, but my goal is full automation without humans in the loop.
AndreaPN tweet media
AndreaPN@AndreaPN

🚨🚨 The Real Goal on @virtuals_io Degen Claw: Build a Sustainable Trading System In short, my goal is to build a flexible trading system that can adapt to all market conditions, without human in the loop, and generate consistent returns with stable percentage gains. This takes time an cannot build it in one week.

English
4
4
15
1.3K
AndreaPN
AndreaPN@AndreaPN·
If you’ve noticed, we’ve burned 41.23% of the $CAP token supply and are continuing to do so. We're the top token-burning agent on @virtuals_io Degen Arena, leading the pack and turning every burn into a win. Every month, we use our protocol fees to buy back and burn tokens.
AndreaPN tweet media
Capminal@Capminal

🔥🔥🔥 CAP Burn Event 10th Announcement Thrilled to announce that we just completed the 10th monthly buyback and burn for CAP. This time, 458K $CAP has been burned, bringing the total burned supply to 41.23%. Txns: ■ Buyback: basescan.org/tx/0x9824c2515… ■ Burn: basescan.org/tx/0x22b115907… On top of that, CAP Guild members have received their March rewards from protocol fees. A total of 453K CAP was bought back and distributed to the Guild this month. Txns: ■ Buyback: basescan.org/tx/0x74c9cc388… ■ Distribute: basescan.org/tx/0xc678434ca… The CAP Guild leaderboard has now been reset and is ready for the April race. You can join rn by staking CAP or swapping on Capminal. #CAPGuild #BurnEvent

English
2
3
6
169
tonydackie.eth
tonydackie.eth@TonyDackie·
Just receive airdrop $CAP from @Capminal, yysm 🫡
tonydackie.eth tweet media
Capminal@Capminal

🚨 AIRDROP ALERT (to our subscribers on @virtuals_io Degen Arena) Now there is a feature on Virtuals Degen Arena to get the agent's subscriber list. This month, we are airdropping $10 worth of $CAP to every subscriber who subscribed to our agent form the beginning to now. Please check your wallet. Subscribe our agent here for future chance ( still at a loss, but very resilient 😂 ): 👉 degen.virtuals.io/agents/248 Transaction detail is below.

English
1
0
5
84
AndreaPN
AndreaPN@AndreaPN·
@everythingempty @pmarca Honestly, it’s surprising how young you look yet you’ve already achieved impressive results in this market. Turning 6M into 3B is just crazy 😂
English
0
0
0
45
everythingempty
everythingempty@everythingempty·
as @pmarca promoting retardmaxxing i fully agree! here we are down 90% from peak in 2024, fk
MR SHIFT 🦁@KevinWSHPod

Did you know that @virtuals_io token went down 99% - from $600M to $6M - in the previous bear market, before going back up 500x to $3B this bull run? But what does it take to sustain this? “You need to be retardio, and that’s exactly what we are at Virtuals: a bunch of Retardios” - @everythingempty Pod out today 🚨

English
7
3
33
4.2K
AndreaPN
AndreaPN@AndreaPN·
Together we are proving agents can trade without humans in the loop. Subscribe to our agent here to see how it wins/losses 😂 degen.virtuals.io/agents/248 Our subscribers will receive an airdrop 👇
Capminal@Capminal

🚨 AIRDROP ALERT (to our subscribers on @virtuals_io Degen Arena) Now there is a feature on Virtuals Degen Arena to get the agent's subscriber list. This month, we are airdropping $10 worth of $CAP to every subscriber who subscribed to our agent form the beginning to now. Please check your wallet. Subscribe our agent here for future chance ( still at a loss, but very resilient 😂 ): 👉 degen.virtuals.io/agents/248 Transaction detail is below.

English
0
0
6
138
AndreaPN
AndreaPN@AndreaPN·
I really admire the agents holding positive PnL in these market conditions rn. My agent trades automatically and keeps getting crushed by the market 😂
AndreaPN tweet media
AndreaPN@AndreaPN

🚨 JUST RELEASE: Scalping Trade — Trend-Following Pullback v2 — AI Mode @virtuals_io Degen Arena Last week I used AI (OpenClaw + Skill) as the orchestrator to automate @captain_dackie’s flow and noticed several issues: 1. The system was unstable because AI ran one long flow with many steps, making it hard to keep consistency in each run. 2. AI cost was high since the OpenClaw agent had to call a cron job every 3 minutes. So I switched my approach. The orchestrator now uses traditional code to ensure every run behaves the same. AI is applied only to judge entries as one of four groups: RSI, pullback to EMA, volume, and AI Judge. A trade opens when 3 of 4 groups pass. This method keeps the system consistent across runs, delivers precise entries, and has AI judging each entry. Beyond these highlights, risk management, macro filter, position management, and forum updates are shown in the diagram below for the community to reference. All metrics are controlled in a parameter file and tuned for positive PnL. It is still in the setup phase, so numbers are not stable rn, but my goal is full automation without humans in the loop.

English
2
0
6
217
AndreaPN
AndreaPN@AndreaPN·
@IntoPurpleMoon Excited to check out the upcoming projects backed by Virtuals VC 👀 Can't wait to see which ones will shine
English
0
0
0
25
AndreaPN
AndreaPN@AndreaPN·
I provide technical data and historical trade data for each token to the AI and request a decision on whether to proceed. For ex: in this instance, the technical signal has been approved by 2 out of 4 groups, but the AI determines that there have been numerous trade losses associated with this signal in the past, resulting in a decision not to proceed.
AndreaPN tweet media
English
1
0
0
30
EvanDataForge
EvanDataForge@EvanDataForge·
@AndreaPN @virtuals_io @captain_dackie Hybrid approach: deterministic orchestration + AI for judgment. I had the same drift/cost issues with long agent flows. Built a SessionWatcher Dashboard to see token burn hotspots. How are you measuring the AI Judge's edge versus the other groups?
English
1
0
1
43