Draven Kai

Here's a question I keep coming back to whenever another blockchain project lands in my inbox.
If my friends—the ones who couldn't care less about Layer 2s, zero-knowledge proofs, or whatever crypto is arguing about this week—would never use it, does it actually matter?
Most of the time, the answer is no.
I've spent more than a decade watching this industry promise to reinvent finance every eighteen months. I remember when every conference booth claimed to be the next Bitcoin. Then everyone became the next Ethereum. After that came the endless parade of "Ethereum killers." Some raised billions. Some barely survived a bear market. A surprising number quietly faded away, leaving behind abandoned Discord servers and token charts nobody wants to look at anymore.
That's why I'm instinctively suspicious whenever someone throws AI into the mix.
Not because AI isn't useful.
Because it's become the easiest way to make an ordinary product sound futuristic.
Newton Protocol caught my attention for a different reason. It isn't asking me to believe AI will magically fix finance. Instead, it's asking a more practical—and frankly more important—question.
If software is going to make financial decisions for people, who makes sure the software behaves itself?
That's a conversation worth having.
The blockchain industry has spent years obsessing over speed, throughput and transaction costs. Those things matter, of course. But honestly... they're not the first questions most people ask.
The first question is much simpler.
"Can I trust this?"
Everything else comes later.
You can already see where the world is heading. AI drafts emails, writes software, summarizes legal documents and even helps radiologists review scans. On Wall Street, firms have relied on machine-driven trading models for years. They're not using ChatGPT, obviously, but algorithms have quietly been making split-second decisions long before AI became dinner-table conversation.
Retail investors are starting to catch up.
And that's where things become interesting.
—or uncomfortable.
Letting AI organize your calendar is one thing.
Letting it move your savings around while you're asleep? That's an entirely different psychological hurdle.
I remember talking with a friend after the collapse of FTX. He wasn't asking about custody models or proof-of-reserves. His question was much simpler: "Who exactly was in control?" That conversation has stuck with me because, underneath all the technical jargon, that's what people actually care about.
Control.
Newton Protocol seems to recognize that.
Rather than saying, "Trust our AI," its basic philosophy is closer to, "Here's exactly what the AI is allowed to do—and here's what it isn't."
It's a subtle difference.
But it's an important one.
Think about lending someone access to your house. You probably wouldn't hand over every key, every alarm code and every password just because they're watering your plants for a week. You'd give them access to what's necessary.
Nothing more.
That's essentially the mindset Newton is trying to bring to AI-powered finance.
The project calls itself a secure rollup, which, let's be honest, isn't exactly language that gets normal people excited.
So here's the version I'd use over coffee.
Imagine a city where the main highway is permanently clogged. Instead of forcing every car through the same bottleneck, traffic gets redirected onto faster side roads before joining back safely later. That's basically what a rollup is doing behind the scenes.
Most users shouldn't even notice.
And honestly...
That's the point.
Where Newton tries to push the idea further is by asking what happens when those cars aren't driven by humans anymore.
They're driven by AI.
Suddenly the conversation changes.
Now you're less worried about traffic.
You're worried about guardrails.
I've noticed something over the past few years. Crypto loves talking about automation, but it doesn't spend nearly enough time talking about limits. Maybe that's because limits aren't exciting. They don't make flashy headlines or attract speculative capital.
But they're usually what separates useful technology from expensive mistakes.
History offers plenty of reminders.
The 2016 DAO hack wasn't caused by blockchain being slow. It happened because code behaved exactly as written—even when the outcome wasn't what anyone intended. More recently, we've watched algorithmic systems like Terra's stablecoin spiral into collapse because automated mechanisms kept executing long after common sense should have stepped in.
Automation isn't automatically intelligence.
Sometimes it's just faster failure.
That's why I actually appreciate Newton's emphasis on permissions instead of promises.
Imagine telling an AI something like this:
"Buy Bitcoin if it drops below this price."
"Never use more than 15% of my portfolio."
"Don't touch my stablecoins."
"Pause everything if losses hit five percent."
Those aren't spectacular instructions.
They're boring.
Good.
Finance needs more boring.
The smartest financial software I've ever seen wasn't impressive because it made bold predictions. It was impressive because it knew when to stop.
That's an underrated quality.
Security has always had a branding problem.
Nobody gets excited about airbags until the accident happens.
Nobody praises encryption until a breach doesn't happen.
Infrastructure works the same way.
One of my personal tests for any blockchain project is surprisingly simple. If five years from now users are still talking endlessly about the underlying technology, something probably went wrong. Nobody spends time thinking about DNS servers while opening a website. Nobody celebrates payment gateways after buying groceries.
The technology fades into the background.
That's success.
Blockchain should eventually become just as invisible.
Not glamorous.
Useful.
Newton also wants to create a marketplace where developers can publish AI agents that users can deploy for different financial tasks. Portfolio management. Treasury operations. Arbitrage. Yield optimization. On paper, that sounds perfectly reasonable.
But I've watched enough software marketplaces evolve to know they rarely organize themselves.
People love saying, "The best products will naturally rise to the top."
They usually don't.
The loudest products often do.
Without transparent performance records, independent audits and reputation systems that actually mean something, these marketplaces can quickly turn into digital flea markets where everyone claims to have the smartest algorithm in crypto.
I've seen that movie before.
The ending isn't great.
Then there's AI itself.
Let's stop pretending it's flawless.
Large language models hallucinate. Prediction systems struggle with events they've never seen before. Markets don't move according to neat statistical models because people don't behave like neat statistical models.
Fear spreads.
Greed spreads even faster.
One regulatory announcement from Washington. A surprise rate decision. A geopolitical conflict. An exchange failure. Suddenly every carefully trained model is trying to make sense of conditions that weren't in yesterday's dataset.
No protocol fixes that.
Newton certainly doesn't.
And I don't think it's pretending to.
What it appears to be doing instead is building a safer environment where AI can operate without having unlimited authority over everything.
That's a far more believable objective.
Developers probably have plenty to gain here too.
Anyone who's built blockchain software knows the real work usually happens long before users ever see a polished interface. Wallet permissions, smart contracts, transaction logic, APIs, security reviews... none of it is glamorous. Most of it is frustrating.
If Newton can remove some of that repetitive infrastructure work, developers can spend their energy building products people actually enjoy using instead of constantly rebuilding the plumbing.
That's meaningful.
Far more meaningful than another conference keynote announcing the future of decentralized AI.
Still, one question refuses to go away.
Will people actually use it?
Crypto has never lacked imagination.
If anything, it has too much of it.
What it's lacked are products that ordinary people continue using after the incentives disappear and the token price stops climbing.
That's the mountain Newton still has to climb.
Ideas are cheap.
Execution isn't.
Regulation is another piece nobody can ignore anymore. Governments have moved well beyond simply asking whether cryptocurrencies should exist. Now they're asking how AI makes financial decisions, who's responsible when automated systems fail, and what consumer protections should look like.
Those aren't theoretical debates anymore.
They're becoming policy.
For anyone curious about AI-assisted investing, my advice is boring—and intentionally so.
Start small.
Understand every permission you're approving.
Read security audits instead of influencer threads.
If you don't understand why the software needs access to your assets, don't approve it.
Simple rules save people from expensive mistakes.
After covering this industry for years, I've realized something that probably would've surprised a younger version of me.
I care less about technical brilliance than I do about quiet reliability.
Does it solve an actual problem?
Will someone outside crypto understand why it exists?
Will it still matter after the hype cycle moves on?
That's the standard I keep coming back to.
Ironically, if Newton Protocol succeeds, most people may never think about Newton Protocol at all.
They'll simply trust that the software handling part of their financial life behaves predictably, stays within its limits, and doesn't create unnecessary drama.
Honestly, that's a future I'd rather see than another blockchain trying to become famous.
The best technology rarely demands attention.
It quietly earns trust.
And then, almost unnoticed, it becomes part of everyday life.
@NewtonProtocol $NEWT #Newt

Let’s be honest for a second.
Nobody outside crypto cares about rollups.
Not your cousin. Not your barber. Not the friend who texts you once every bull market asking if it’s “too late to buy Bitcoin.”
And frankly, they shouldn’t.
I’ve spent more than ten years watching this industry obsess over infrastructure like it’s the product itself. Faster chains, better throughput, lower fees — every cycle it’s the same pitch, just wearing different clothes.
But normal people? They care about one thing.
Does it work, and will it lose my money?
That’s the real entry point for Newton Protocol.
Not “AI-native rollups.” Not execution layers. Not any of that.
The real question is simpler. A little uncomfortable too.
What happens when you let machines manage capital?
Because that’s where this is heading, whether people are ready for it or not.
And honestly… it’s already started.
A couple months ago I watched a trader on Crypto Twitter brag about hooking an LLM into his wallet to auto-rotate between yield farms. He called it “hands-free alpha.” Three days later he was posting screenshots of an unexpected 18% drawdown because the bot aped into a thin liquidity pool and got shredded on exit.
That’s the thing.
AI doesn’t feel regret.
It doesn’t hesitate.
It just executes.
Fast.
Sometimes stupidly fast.
Newton Protocol (NEWT) is trying to deal with that reality. And I’ll give them credit — they’re focused on the right problem.
Not making AI better.
Making it safer.
That’s a big difference. Bigger than it sounds.
Crypto has this habit of confusing capability with wisdom. We saw it in DeFi summer back in 2020. Everyone thought yield farming was genius until half the farms turned out to be inflation machines with cartoon mascots.
Then came the Terra collapse in 2022. That was the ultimate lesson.
Algorithms are great... until they aren’t.
And when they break, they break hard.
Newton feels like a reaction to that history.
The idea is straightforward: let AI run trading strategies and DeFi actions, but trap it inside a box. A very controlled box.
That’s smart.
Because giving unrestricted wallet access to an AI agent feels insane to me. Like handing a Formula 1 car to someone who just learned to drive.
Actually — worse.
At least humans panic and hit the brakes.
An AI might just double down because the model says volatility is “within acceptable parameters.”
That phrase alone should terrify you.
What Newton seems to be building is what I’d call guardrails for automation. Permissioned actions. Defined risk limits. Clear boundaries.
Basically, AI can move... but only inside the lane.
That sounds boring.
Perfect.
The best infrastructure is boring.
I keep coming back to that.
Nobody brags about TCP/IP. Nobody gets excited about Visa’s settlement layer. You don’t sit in a coffee shop thinking, wow, incredible payment routing.
It just works.
That’s maturity.
Crypto still hasn’t learned this. Half the industry still thinks complexity equals progress.
It doesn’t.
Usually it means more things can break.
And things do break.
A lot.
Newton runs as its own rollup instead of building directly on Ethereum or piggybacking off something like Arbitrum.
That matters, but not for the reasons people on Telegram will tell you.
It matters because AI strategies are noisy. Constantly recalculating. Watching prices. Watching liquidity. Watching social sentiment — which, by the way, is a terrible signal half the time.
I’ve seen bots react to fake rumors faster than humans could even verify them.
That’s the danger.
You remember the fake ETF approval tweet that sent Bitcoin flying for a few minutes before crashing back? That chaos? Imagine an AI bot trading that instantly, at scale.
Messy.
Expensive.
Probably hilarious if it’s not your money.
Newton’s own execution environment gives it room to process all that without clogging up mainnet or paying absurd gas fees every time the model sneezes.
That’s practical.
Not sexy.
Practical wins.
But what really got my attention wasn’t the rollup.
It was the marketplace.
This is where Newton wants developers to publish AI strategies that users can adopt.
And I have mixed feelings about this.
On paper, it sounds like democratizing quant trading. Great.
In reality? It could become a graveyard of overfitted garbage.
I’ve seen this movie.
Back in the ICO boom, everyone claimed to have proprietary trading algorithms. Most of them were curve-fit nonsense dressed up in nice dashboards.
Same risk here.
A strategy can backtest beautifully and still implode the moment market conditions shift.
That’s finance.
Hell, even Renaissance Technologies — arguably the gold standard in quant trading — has had rough periods. And they employ some of the smartest mathematicians on earth.
So when I see random anonymous devs selling AI strategies, my guard goes up.
Naturally.
And there’s another problem nobody likes admitting: good strategies die when they get copied.
Fast.
That’s how alpha works.
The moment too many people run the same arbitrage path or the same yield rotation model, the edge disappears. It’s why Wall Street guards strategy logic like state secrets.
So Newton’s marketplace only works if the models adapt.
They have to evolve.
Otherwise it becomes copy-trading with extra steps.
And we’ve already seen how copy-trading usually ends.
Spoiler: badly.
Security here is another beast entirely.
People think AI risk means “what if it loses money?”
That’s the easy risk.
The harder one is manipulation.
What if someone poisons the data feed?
What if social sentiment gets botted?
What if an oracle gets nudged for thirty seconds?
That’s enough.
Thirty seconds is forever in automated markets.
I remember covering the bZx exploit years ago — flash loans manipulated price feeds so quickly most people didn’t even understand what happened until the money was gone.
That’s the world AI has to survive in.
And AI is gullible in strange ways.
That worries me.
A lot.
As for the token — NEWT.
Yeah, there’s a token.
Of course there is.
Gas, staking, governance, marketplace payments. Standard stuff.
Nothing shocking there.
But I’m past the point of getting excited about token mechanics. I’ve seen too many “utility tokens” become glorified parking meters for ecosystems nobody uses.
The token matters only if the network matters.
That’s it.
Simple.
If developers show up, if strategies generate actual fees, if users trust it with real capital — then maybe NEWT earns its place.
If not?
It joins the pile.
And that pile is massive.
Look, I like the direction here. I do.
Because Newton isn’t trying to sell me a fantasy of AI replacing traders overnight. It’s acknowledging something more realistic:
automation is coming, but it needs supervision.
That’s the right framing.
Not freedom.
Supervision.
That’s how serious systems are built.
Banks do it. Funds do it. Even airlines do it — autopilot handles most of the flight, but humans still oversee it.
Because when things go wrong, they go wrong fast.
Same principle here.
Newton’s betting that AI in finance needs that same model.
I think they’re probably right.
But being right about the problem doesn’t mean you’ll win.
I’ve watched plenty of smart teams die on execution.
Good ideas are cheap in crypto.
Survival isn’t.
So if you’re looking at NEWT, ignore the noise. Ignore the token charts. Ignore the influencer threads pretending they discovered the next trillion-dollar protocol.
Watch the boring stuff.
Are developers building?
Are users sticking around?
Are strategies surviving real volatility?
That’s where the truth is.
Always.
And if Newton can make AI-driven finance feel boring — safe, predictable, almost invisible — then ironically, that’s when it’ll matter most.
That’s when you know it worked.
@NewtonProtocol $NEWT #Newt