SPACEWHALES
1.9K posts
SPACEWHALES
@D2HLC
grinding the trenches for fun and profit
Katılım Ekim 2013
894 Takip Edilen172 Takipçiler
Following a weekend of expanded usage allowances, Anthropic’s highest-tier subscribers are waking up to a crippling rate-limit bug.
Developers paying top dollar for the "Claude Max" ($100/mo) and "Max 20x" ($200/mo) plans are reporting that their accounts are being locked out almost instantly due to an issue with how Claude Code is calculating token consumption.
Brad Groux@BradGroux
Something is up with Claude Code usage today. $200 Claude Max, 0%, 52% to 62%, then 68%, 76% and 84% in 5-hour rolling window in the time it took me to write this tweet. WTF, @AnthropicAI? I'm working on one GitHub PR for regression testing. Not folding proteins to cure cancer.
English
Notes I kept from the recent 1000x podcast with Capital Flows:
We're in the golden age of discretionary trading. Individual traders can move fast & embrace volatility that big funds can't. That's your edge. Use it.
True edge isn't just price. Pure TA is a fool's errand. Layer in macro data, global liquidity flows & central bank certainty or you're flying blind.
Control your information diet. Read sell-side research to find holes in your own thesis. Long-form macro in the morning. Social media = headlines only, never research.
Geopolitical wildcard: A US-led regime change in Iran could cripple China's global apparatus & trigger a massive rush back into US equities, killing gold & the multipolar trade overnight.
Commodity play: Western uranium & rare earth miners. Nuclear is non-negotiable for powering AI data centers. Avoid anything China-supply-chain dependent.
High-conviction 2026 trades:→ $PER best proxy for Hyperliquid's TradFi/crypto bridge → Oracle. Larry Ellison is betting the entire balance sheet on AI infra. Exponential upside.
English
@aakashgupta How does this affect the weekly limit? I don't want to just reach it faster
English
This two-week promo tells you more about AI infrastructure economics than any earnings call this quarter.
Anthropic hit $19 billion in annualized revenue as of early March. They raised $30 billion at a $380 billion valuation five weeks ago. Claude Code alone crossed $2.5 billion annualized, doubled since January.
And they have so much spare GPU capacity outside a six-hour weekday window that they can double usage for free across every tier.
That’s the demand curve problem. Peak hours (8am to 2pm ET) consume the infrastructure. The other 18 hours per day, plus all of Saturday and Sunday, the GPUs sit underutilized. You’re paying for hardware sized to handle Monday at 10am, then watching it idle at 9pm.
Run the actual numbers. Anthropic expects to spend roughly $7 billion on inference this year and another $12 billion training models. If utilization outside peak drops even 40%, billions in capacity sits dormant for 75% of the week.
Electricity companies solved this decades ago. Time-of-use rates, off-peak discounts, demand response programs. Fixed infrastructure, variable demand, massive waste during idle periods. The economics are identical.
Two weeks of doubled off-peak limits generates exactly the behavioral data Anthropic needs: does free capacity actually pull users into evenings and weekends? If it does, they just discovered a lever that smooths GPU demand without building a single new data center.
The AI lab that cracks utilization shaping first gets the equivalent of a free fleet of H100s. Anthropic just started the clock.
Claude@claudeai
A small thank you to everyone using Claude: We’re doubling usage outside our peak hours for the next two weeks.
English
@ieaturfoods Yeah still holding that sh*t since a YouTube charlatan shilled it 4 years ago (ends with aiki)
English
NEAR is quite literally the definition of an unserious project and hops from one popular narrative to the next fwiw… just incase people forgot
-general purpose L1
-gaming partnerships
-algo stablecoin
-AI focused layer 1
-privacy focused L1
-AI/“shade” agents blockchain
Stablecoin Sean@seanlippel
$near is what everyone hopped that $tao would be as far as a leading blockchain for AI i've long said this (mostly bc of the fundamentals issues with the tao model), but now the market is realizing and quickly repricing both i think near flips bittensor for good by month end
English
@DonAlt You can always tell when the next duck has been through their shit on life and became the bigger person
English
>dad died when I was 12 yo
>mom had to raise me and my brothers alone
>grew up poor
>did decent in school
>went to university
>got cancer
>recovered
>started trading instead of finishing university
>decent at it
>not poor anymore
>met wife
>two beautiful children
>no sleep
✧@cessonmute
give me your survival arc
English
Day 440: Polymarket just introduced “Sponsor Rewards.”
Users can now deposit USDC to reward liquidity providers and incentivize deeper liquidity in specific markets.
Since Polymarket keeps rewarding users who provide liquidity, I’m fairly confident this will become an important multiplier for the $POLY airdrop.
If you don’t know what to do in this boring market, dive deeper into LP rewards and consider allocating some capital to Sponsor Rewards as well.
Position yourself accordingly
Mustafa@mustafap0ly
sponsoring market rewards is now open to all users 😛 add rewards to any market to get the liquidity for the size you want to trade. permissionless market deployment and creator fees next...
English
UPDATE: $480 left to go.. about 3 hours left on the clock for $10k in a week
Pray for me 🙏
Xeny@XenyPM
Almost at $10k for this week... I have a few hours before my initial profits this week are discarded Will I make it before the deadline? We'll find out shortly
English
@Trader_XO @thdxr Could you share any quality tutorials? What is your main use case?
English
@thdxr Shared this earlier.
x.com/Trader_XO/stat…
XO@Trader_XO
As someone who leverages both Claude Code and Codex (with a preference towards Codex) - theres a plethora of info, blogs, videos, newsletters geared around Claude Code but not so much Codex. Why....? Groks reply... FWIW both are fab and I run both simulatanously
English
codex is by far a better coding model than opus - anyone who knows anything understands this
but the whole industry should reflect on why opus is the most popular
people assume whatever is the smartest will win but the old rules of product are still what determine everything
English
@YujungHwang3 I'm interested to know which vscode extensions you use other than cc. I find positron a more familiar environment
English
Think it's time to update my mega $HYPE DCA journey,
13 months (405 days) after the beginning of it, not at peak euphoria ATHs, but in the middle of the bear market
For those who don't remember or know how this worked, for the past year I kept consistently accumulating spot $HYPE on Hyperliquid through a RSI-based strategy
The strategy was getting fed $50 per day and it bought based on the daily RSI:
RSI => 70 don't buy
RSI<= 30 buys with 100%
RSI between 30 and 70 buys with portions always smaller the higher the RSI goes
This strategy ensured a big advantage against a classic flat DCA
Let's see the numbers
Total $ spent: $19.965$
Total $HYPE accumulated: 723,24
Average dollar cost: $27,6
Even after a big overall market correction, the DCA is in profit at the time of writing this
On the contrary, a classic flat DCA would've gotten a dollar cost of $31.42
This means we get 11% APY of more HYPE from a flat DCA strategy
Year 1 has gone, 7 more to go
Pasta Capital@Pasta_Capital
I am starting a RSI-based Dollar Cost Averaging on $HYPE for 50$ daily until the NEXT 2 CYCLES (likely and usually 8 years) Using a fresh wallet in order to track it perfectly More info 👇🧵
English
@ReetikaTrades It's a nice boost when you get a nice score though.
Works both ways
English
Let's talk timelines on this. The amount of time it takes for a pattern to break out is typically proportional to the time frame the pattern forms on. Given this is a relatively high time frame setup(8M), this setup could take months or even years to breakout, assuming it's not invalidated in the meantime.
If the 4 year cycle remains alive, that could line up quite well with a breakout during the next bull market (2027-2029). If the 4 year cycle is dead then everything is on the table including a possible global recession dragging the next crypto bull market out beyond 4 years, including this setup.
Not a "quick buck" type of trade by any means.
English
$ETH/BTC has one of the most bullish setups I've ever seen.
The flippening is real 🐬
English
The author of Clean Code is using Claude to write software.
I can hardly think of a clearer indicator that coding is now officially outsourced to LLM’s.
Who else is mourning the death of coding by hand ?
Uncle Bob Martin@unclebobmartin
I'm keeping two Claude windows open. One making changes to the source code. The other helping me to plan future changes. This uses my time better since I spend a lot less time waiting for Claude to finish. The two Claudes are in two different directories, both have git repos of the project. The planning directory has rules that prevent source code changes. Only planning is allowed. I manually pull and push changes between the two.
English
@sankitdev @antigravity yea no point anymore...
feels like the jump from assembly -> c -> higher lvl languages
only much bigger
English
@D2HLC @antigravity Tbh its not lost but repeated stuff should be automated. Nobody should focus on those Ui issue thats what i feel.
English
huge, Codex won't be like dry bread to talk to anymore, which was one of the strengths of Claude Code
Kevin Kern@kevinkern
Codex 0.92 has a new option /personalities. friendly: longer, conversational pragmatic: short and explicit Right now, it looks hardcoded, so we can't create custom personalities. Still, it's a nice addition.
English
@thesamuelnam No problems on m2 base model, using it all day for data science and web dev
English
Am I the only one on the M1 chip avoiding the update like my life depends on it?
English