Sabitlenmiş Tweet
jay
730 posts
jay
@jayqxr
blockchain analyst||shitcoin trader
on-chain Katılım Ağustos 2020
33 Takip Edilen32 Takipçiler
jay retweetledi
when i got my first job in crypto, i was preparing for one of my backlog engineering exams.
since then i’ve seen incredibly capable people unemployed, while others with far less understanding sit in top firms.
startups should bet more on people who are hungry to learn and build, not just degrees, countries, or referrals.
and if crypto hiring runs on reputation and online presence, then be hungry enough to build that too.
too diabolical but it’s always 2 ways.
Brian Armstrong@brian_armstrong
Some of our best hires were totally unqualified on paper. They always had the same qualities: entrepreneurial, high agency, smart, mission aligned, and they got shit done. If you’re hiring, especially in early stages, seek out & bet on these people. Don’t over-index on resumes.
English
@jayqxr DM me on how this would help @rallyguardco in plain english and make it actionable for the developers and bizdev people.
English
I think I’ve applied to about 20-30 crypto related roles over the last 2-3 months and not a single one even wants to interview me
I graduated college and am decently knowledgeable about this space
It’s all I’ve done since I graduated
I think it might be time to give up
English
@fuhadd_sol Hey, what kind of roles are you applying for? I have a project maybe it could be a bit of resume builder for you. DM, also maybe their is someone in my network looking to hire.
English
jay retweetledi
There’s a phase nobody warns you about
It’s not the crash. It’s not the rug. It’s not even the losing streak
It’s the nothing
The weeks where everything feels slow
The setups are there but nothing’s hitting
Just you staring at screens wondering if you’re wasting your time
That phase isn’t a sign you’re failing
It’s where everyone who wasn’t serious falls off naturally
If you can survive the boring, the exciting part takes care of itself
English
jay retweetledi
jay retweetledi
A lot of people saying SOL today feels like it did at $8. What a wild thing to be saying.
At $8, we were all concerned that the chain was actually dead:
- FTX held a substantial amount of SOL.
- FTX had been hacked.
- $8 SOL made a 51% attack realistic.
- Every bridged asset was depegged.
- Projects went from 8/9 figure treasuries to zero.
- Users had lost their life savings.
We were standing on deaths doorstep.
Today:
- The chain has had no issues in a long time.
- The ecosystem is thriving.
- Chain revenue is the highest in all of crypto.
- Institutions are here.
- RWAs are growing.
Crypto is cyclical. We will be back again like we have never been back before.
@solana has never been better positioned.
English
jay retweetledi
Memecoin trading isn’t gambling, it’s skill
Show me any other market where a repeatable charting structure hits 6-8/10 times in high liquidity times, lets you buy spot at structural lows, risk 20-30% downside on invalidation, and consistently (skill) prints 50-500%+ upside.
Only memes.
That asymmetry doesn’t exist anywhere else.
Not equities. Not forex. Not commodities. Not majors.
Memes are the only market where the structure is respected even with high volatility, liquidity is reflexive, narratives amplify technical breakouts, spot entries at lows routinely outperform leverage, failure cost is capped and winners are outsized.
Meme coin trading isn’t gambling, it’s structural inefficiency.
Memecoin markets are structurally inefficient because they are dominated by emotional retail participants whose narrative driven behavior repeatedly misprices structure and creates exploitable asymmetry.
How to be profitable?
- disciplined mindset (read Psychology of Money, trade with small size)
- repeatable strategies (find your edge, stick to it, boring? its work, not your vacation)
- established routine (fixed trading rhythm each day, spot the best one over time)
- simple TA (use a few basic patterns, TA is human psychology and memes are pure emotions)
- create trading rules (if...then...)
- mainfest your trading principles (unshakable basis)
- learn to avoid emotional decision making (keep your size small, limit the number of weekly trades...)
- use a trading platform like Axiom (tools will give you advantage, clean UI, focus on one screen...)
Also important, and often underestimated, is exchanging ideas with other traders.
Pay attention to what’s happening on your timeline, read actively, and educate yourself in real time by absorbing what others are sharing on X.
Many traders who have survived multiple market cycles openly share their experience on the timeline. Not everything needs to be copied or acted on, but a lot of it compounds subconsciously.
That’s how you build your own training framework over time: by continuously observing, filtering, and internalizing high quality market thinking.
English
jay retweetledi
People ask me what the secret is
There isn’t one
It’s risk management. It’s patience. It’s cutting losses fast and letting winners run. It’s not overtrading. It’s not sizing up on emotion
Boring stuff. Stuff everyone knows
The secret isn’t knowing it. It’s doing it consistently when everything in you wants to do the opposite
That’s it
English
The beanstalk is sprouting stronger! ✨
Teaming up with Gassy Jack to climb the beanstalk. Collect beans through quests and unlock epic rewards in our gas-free world. Who's in? 👀
ethgas.com/community/onbo…
ethgas.com/community/onbo…
English
@defi__josh @Ronin_Network @Jihoz_Axie @AxieAur @Dune @Dune_Africa @arno39 @smartymetrics @0xTobey @_muhraf_ amazing work josh. can’t wait to reach this level 🤞
English
Gm Gm..
I’ve built and shipped machine learning systems in production, not just models.
Both in large-scale Solana game analytics and ML-Powered Churn Prediction x.com/i/status/19983… and in my latest build, ronhub.io/analytics.
And one thing has been consistent every time:
The model was never the hard part.
On ronhub.io, ML only works because the system around it works.
Before touching any model, I designed the full data pipeline:
• Raw onchain ingestion
• Cross-sector normalization
• Deterministic feature generation
• Aggressive caching and monitoring
• Reproducible transformations
That same structure came from my earlier Solana games work, where I processed 60M+ transactions and built churn prediction and behavior forecasting directly into the pipeline.
No notebooks running in isolation.
No manual experiments.
No “rerun and hope”.
Data flows in.
Features are created automatically.
Models retrain on schedule.
Predictions are evaluated continuously.
Failures are visible immediately.
At that point, model choice becomes secondary.
What actually matters is:
• Can the pipeline survive scale?
• Can you detect drift early?
• Can you swap models without breaking the product?
• Can teams trust the outputs?
That’s how ML exists inside RonHub today.
Not as a headline feature.
But as an internal system that strengthens health scoring, behavior analysis, and ecosystem monitoring on Ronin.
This is why I don’t think in terms of “building models” anymore.
I build systems where models are expected to change.
Model building is table stakes now.
Owning the system around it is where leverage compounds.
If you’re serious about ML, learn how to design pipelines first.
Models will follow.
Jo$h@defi__josh
You know that feeling when you're analyzing blockchain data and you realize... game developers are flying blind? 😳 Millions of players. Billions in transactions. Zero tools to predict who's about to leave, etc.. That hit me hard while diving into @solanagaming ecosystem. Because here's the truth: 👉 By the time a player churns, it's already too late. So instead of just tracking what already happened alone... I built something that predicts what's about to happen. =============================== 🎮 Solana Game Analytics, Player Behavior Modeling & Predictive Forecasting (ML-Powered Churn Prediction) =============================== A production-grade platform that analyzes 60M+ transactions from 12 @Solana games ( @AuroryProject, @staratlas, @genopets, @Stepnofficial, @PlayHoneyland, @maxxmxmcoin, @RIFTAI_, @play_evio, @MiniNations, @_portals_, @nyanheroes, @MixMobRacer1), predicts which players will churn 14 days in advance with >85% accuracy, and auto-retrains its ML models whenever fresh blockchain data arrives. .... I love the sound of that BTW 😊 --- 1️⃣ What It Does ↘︎ Aggregates 60M+ on-chain transactions from 12 major Solana games ↘︎ Predicts player churn 14 days before it happens (>85% ROC-AUC accuracy) ↘︎ Auto-retrains 5 ML models and selects the champion automatically ↘︎ Tracks 11 behavioral metrics: activation, retention, cross-game behavior, etc. ↘︎ Delivers insights through 21 REST API endpoints + a gamified real-time dashboard. ..... Tbh. .. I love this UI so much 🥰 ..... Everything updates automatically. Everything is backed by verified Solana blockchain data...with clickable links. --- 2️⃣ How I Built It ↘︎ Data: 11 custom @Dune Analytics queries + custom classifier Python Script for on-chain analysis ↘︎ ML Pipeline: 5 algorithms (Random Forest, XGBoost, LightGBM, Gradient Boosting and Logistic Regression) trained in parallel with SMOTE balancing + auto-champion selection ↘︎ Backend: @FastAPI (21 endpoints, 72-hour caching, sub-100ms responses) deployed on @Railway ↘︎ Frontend (My favourite 😍): @nextjs + React 19 (100% TypeScript, Solana-themed UI, 30s auto-refresh) on @vercel ↘︎ Features Engineered: 10 predictive signals per user — activity patterns, momentum, consistency, recency --- 3️⃣ Why This Matters Solana gaming is exploding... but lack retention tools. This dApp (has Wallet Connect feature) gives: 📍 Early warning signals — catch at-risk players 2 weeks before they leave 📍 Cross-game insights — understand behavior across the entire ecosystem 📍 Zero-maintenance ML — models auto-improve with new data 📍 Actionable intelligence — trigger campaigns, interventions, VIP monitoring For Solana: 📍 First comprehensive gaming analytics platform 📍 Open-source models + all 11 Dune queries for community use --- 🔗 Explore the System • Full Demo: drive.google.com/file/d/1Pr1j8z… Built by: @defi__josh Inspired by: @indiedotfun @oladeeayo @DairusOkoh @apostleoffin @AnalyticSages @alphabot @superteam @KttyLabs @Purrvananew --- Fr... This one fused blockchain analytics, Machine Learning, full-stack engineering, and DevOps into a single autonomous system. What excites me most? 👉 Game studios can now predict churn before it happens. 👉 The system retrains itself — zero manual work. 👉 The UI is bada_ss 😊😇 Because if Web3 gaming is going to scale, we need tools that help developers keep the players they earn. 💯 --- BTW.. 👉 What's the #1 reason you think players churn from blockchain games? --- Gm Gm 😉
English
