Jonathan Bakebwa
3.4K posts

Jonathan Bakebwa
@codebender828
Building @DonutAI // Co-Founder @SonicSVM // Born of God⚡🇺🇬 🇨🇳 🇺🇲
Earth Katılım Eylül 2011
2.7K Takip Edilen3.3K Takipçiler
Sabitlenmiş Tweet

@nhanvp @mert @nick_pennie @japarjam @toly True. 👀 So some product-oriented questions wouldn't be too far off from target.
English

@codebender828 @mert @nick_pennie @japarjam @toly The disciplines of eng, product and design are merging. Everyone thought PMs were gonna become eng too but the reverse seems to be happening. As such:
Staff/principal level only. Strong product/business sense, should be able to run a tech or product line independently
English

@nick_pennie @mert @japarjam @nhanvp @toly Good note.
Is the coding round peered (are you on the call with them?) Been giving take home exercises, and for each test, I have a passing criteria and I look for some things that would demonstrate depth of understanding, ownership and digging deep.
Considering taking it sync.
English
Jonathan Bakebwa retweetledi
Jonathan Bakebwa retweetledi
Jonathan Bakebwa retweetledi

1/ SIMD-0321, authored by @realbuffalojoe and @deanmlittle, provides a direct pointer to instruction data in VM register 2 at program entrypoint. Programs no longer need to parse through all accounts in the serialized input just to reach instruction data 🧵
English

To be fair though most apps actually glorified forms.
Dmitriy Kovalenko@neogoose_btw
Dude has been vibecoding 10k lines a day burnt 100m tokens to built … an email newsletter 🥴?
English

Just met cool guy @alanssol for dinner. Cool catch ups as always.
English
Jonathan Bakebwa retweetledi

1/ A few thoughts on where we're headed with personal agents in trading (and why most trading apps will die): 🧵
2/ It's easier than ever to build a trading app with AI. But most will disappear. Instead, we'll delegate trading tasks to personal agents that work across exchanges, data sources, and execution layers.
3/ The idea that traders will use 100s of tools — TradingView, Coinglass, Telegram bots, Discord alerts, 5 exchange apps — is a dead end. Context switching kills edge. Instead: ONE agent that knows your strategy, risk tolerance, and handles the busywork.
4/ The trading UX we know — dashboards, chart tabs, alert configs — will fade away. The ultimate UX:
→ "What's the funding rate arb opportunity right now?"
→ "Set a martingale grid from 60k to 55k"
→ "Close my SOL if funding flips negative"
No clicking. Just outcomes.
5/ This is what we're building at @DonutAI. D0 CLI is the execution layer — 100+ commands across trading, market data, macro, and DeFi:
d0 hl positions
d0 macro fear-greed
d0 liquidations --asset BTC
d0 order --asset BTC --side buy --size 0.1
6/ But CLI is just the foundation. The real product is the agent layer on top:
→ Persistent memory of your trading history
→ Learns your risk preferences
→ Executes across any venue
→ Knows when to ask vs when to act
7/ The requirements to build a great trading agent = hiring a great trader:
→ Executes fast when opportunities appear
→ Manages risk even when you're emotional
→ Tracks what's working, what's not
→ Available 24/7 across any device
8/ Nobody has built this end-to-end yet. Trading bots execute but don't learn. AI chatbots analyze but can't execute. Dashboards show data but require you to act. The agent that does all of it? That's the gap.
9/ What we're building:
D0 CLI — Command layer (execution)
WebApp — Trust layer (verification)
Agent — Intelligence layer (compounds)
CLI = power. WebApp = trust. Agent = leverage.
10/ The future of trading isn't "AI trading apps." It's one agent that knows your edge, executes your strategy, learns from every trade, handles the 90% so you focus on the 10%.
11/ If you're a trader frustrated with tool fragmentation, or a builder exploring this space — DM me. We're looking for early users who want to shape what trading agents become.
English

Jonathan Bakebwa retweetledi
Jonathan Bakebwa retweetledi

Going manic providing institutional grade price data for @ManicTrade’s high frequency markets 🔮
Manic Trade@ManicTrade
Manic × @PythNetwork Manic integrates Pyth to stream live price feeds for our High-Frequency Prediction Markets. From $BTC to Gold, Stocks, and Forex, predict the move of any Global Asset across 30s - 5m timeframes. ️ Powered by the fastest oracles on Solana.
English







