TraderLi

Quant trader, macro researcher, crypto and NFT investor

[Khairallah AL-Awady @eng_khairallah1]

Boris Cherny, the creator of Claude Code at Anthropic, just explained why single-agent workflows are already dead

in this talk he breaks down exactly how the future is teams of agents, not better prompts:

- the 14% you lose to CLAUDE.md before typing a word
- one agent researching. one building. one reviewing. one orchestrating
- the architecture that separates hobbyists from real builders
- the 3 properties every agent team needs to actually survive

if you've been using Claude for more than a month and never left the chat window, you've been using one agent when you could be running a team of them

instead of another show tonight, watch this

make sure to bookmark it before it gets lost in your feed

the guide is in the article below

[darkzodchi @zodchiii]

Anthropic engineer showed how one person can run 5 AI agents, that code, test, review, and deploy at the same time.

In 30 minutes they built the whole thing live in one session.

Here's what they cover:

> when to use one agent vs a full team
> how to split work so agents don't step on each other > the exact framework for deciding what each agent handles

that's exactly why, I put together a guide on building agent teams that actually work.

full guide in the article below 👇

[Andrej Karpathy @karpathy]

Wow, this tweet went very viral!

I wanted share a possibly slightly improved version of the tweet in an "idea file". The idea of the idea file is that in this era of LLM agents, there is less of a point/need of sharing the specific code/app, you just share the idea, then the other person's agent customizes & builds it for your specific needs.

So here's the idea in a gist format: https://t.co/NlAfEJjtJV

You can give this to your agent and it can build you your own LLM wiki and guide you on how to use it etc. It's intentionally kept a little bit abstract/vague because there are so many directions to take this in. And ofc, people can adjust the idea or contribute their own in the Discussion which is cool.

[Andrej Karpathy @karpathy]

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

[Anthropic @AnthropicAI]

New Anthropic research: Teaching Claude why.

Last year we reported that, under certain experimental conditions, Claude 4 would blackmail users.

Since then, we’ve completely eliminated this behavior. How?

[ClaudeDevs @ClaudeDevs]

Two updates to auto mode:

· Now available on the Pro plan
· Sonnet 4.6 is now supported, alongside Opus 4.7

Shift+tab, and let Claude run.

[Tobias Reisner @reisnertobias]

One of my favorite features of Hyperliquid is the Vaults.

Most people would be better off depositing USDC in a good vault than trading themselves.

Saves a lot of stress, time and pays better.

Downside:
Doesn't give as much dopamine as sweating on your own bad trades.

Check @SystemicStratHL if interested.

Not sponsored, I am just fan of their work.

[Claude @claudeai]

Tinkering, prototyping, and seeing what happens with Claude Design:

[Claude @claudeai]

Kay Zhu is the co-founder and CTO of @genspark_ai, the all-in-one AI workspace built on Claude.

In a market moving this fast, where anyone can build, he thinks the team is what makes the difference:

[Anthropic @AnthropicAI]

We’re donating Petri, our open-source alignment tool, to @meridianlabs_ai, so its development can continue independently.

Working with Meridian Labs, we’ve also released a major update that improves the adaptability, realism, and depth of Petri’s tests.

https://t.co/CyicsIScJi

[Anthropic @AnthropicAI]

Last month we launched Project Glasswing, our collaborative AI cybersecurity initiative. Since then, we and our partners have found more than ten thousand high- or critical-severity vulnerabilities in essential software.

[Zain Shah @zan2434]

Imagine every pixel on your screen, streamed live directly from a model. No HTML, no layout engine, no code. Just exactly what you want to see.

@eddiejiao_obj, @drewocarr and I built a prototype to see how this could actually work, and set out to make it real. We're calling it Flipbook. (1/5)

[Elon Musk @elonmusk]

Calling him “Scam” Altman is accurate.

This is very much worth reading.

[Hyperliquid Traders @hyperdashtrades]

In just 48 hours, the wallet 0x6355 linked to @FlowTraders deposited a total of $21.3M in USDC on Hyperliquid then transferred the full amount to a secondary wallet 0x3037

Activity on the second wallet shows a significant expansion across multiple positions most notably on Crude Oil where exposure has been increased to $73.6M notional making it one of the largest oil positions on Hyperliquid

The strategy appears to span multiple assets with both long and short exposure consistent with market-making behavior

Currently, the second wallet holds approximately $46.1M in account value

[Hyperdash @hypurrdash]

UPDATE: All three insider wallets have now been intentionally liquidated.

HLP currently holds a $13M notional long position on FARTCOIN, sitting at an unrealized loss of $696K.

[MLM @mlmabc]

HLP is currently long 76M Fartcoin (~$15.4M) and is already down ~$1M on the position.

[merp @0xMerp]

TLDR
between apr 8 - 15 the xyz perp will compress by $14 (if nothing changes)
this is 13%
in theory funding needs to be -0.069% /hr or lower for longs to breakeven

we will see what ends up happening

[Yaugourt.hl @Yaugourt]

HIP-4 isn't just prediction markets. It's an options engine.

The binary outcome format (YES/NO settling at 0 or 1) is the building block for replicating TradFi options strategies natively on Hyperliquid L1.

Example: synthetic call spread.

A deployer creates multiple binary markets on BTC with the same expiry but different strikes:

BTC above 68K? YES = 0.80
BTC above 69K? YES = 0.65
BTC above 70K? YES = 0.50
BTC above 71K? YES = 0.35
BTC above 72K? YES = 0.20

You buy YES on all five. If BTC finishes at 71,500, the first four pay 1.00, the last pays 0. Your payoff increases with every strike BTC clears. The tighter the strikes, the closer it approximates a vanilla call with linear payoff.

Now imagine a frontend that abstracts this. The user sees "Buy BTC Call, strike 68K, expiry March 27". Behind the scenes it's buying a basket of HIP-4 binary outcomes at different strikes. One click, synthetic option.

What else can be built:

Covered calls: long BTC perp + sell YES on "BTC above 75K". Keep the premium if BTC stays below.

Protective puts: long BTC perp + buy NO on "BTC above 65K". Downside protection if BTC crashes.

Straddles: buy YES on "BTC above 72K" + NO on "BTC above 68K". Bet on volatility, not direction.

Capital-protected notes: 90% in yield (lending/staking), 10% in YES tokens. Can't lose more than 10%.

Range accumulators: sell at both extremes, collect premium while BTC stays in a range.

All of this on the same L1 as perps, spot, lending, and staking. One margin account. Full composability.

Hyperliquid.

[Hyperdash @hypurrdash]

1/ About 3 weeks ago, a wallet withdrew $1.32M USDC from Bitget. One day ago, the funds were split across six wallets

Between 15:00–17:00 CET, the entity began aggressively longing $XPL placing buy orders at avg entry of $10.95

The combined long exposure across the six wallets grew to approximately $14.3M

This cluster generated +$2.3M in PnL before the entire position was forcibly liquidated with HLP taking over the exposure

Following the takeover, $XPL dumped further, resulting in a loss for HLP