0xBi

I've seen many say that Claude Cowork is going to cancel GPT. I probably will keep my GPT subscription. There's only one reason: my current arbitrage code can only be modified with Codex in conjunction with GPT-5.2-high. Other models can be patched and fixed, but major rewrites are likely to introduce all sorts of bugs, and only GPT-5.2-high can solve them.

Finding ways to maintain a complete context about oneself will become increasingly important.
Talking with different AIs about different topics always requires a way to consolidate all the key information into one place, a place that can be accessed by all large models and continuously iterated upon as new context input, to achieve higher quality responses in future interactions.
Obsidian is one such tool. Are there better tools or third-party services available?

Used GLM-4.7, quite user-friendly, and most importantly, it is much cheaper than GPT 5 and Sonnet 4.5. The performance feels not much different, probably around 80%. It is quite friendly for someone like me who needs the AI to repeatedly reflect and check, and to perform overall code reviews frequently. For difficult problems, I still use GPT5.2 High and Claude Opus.

Let's review all the trading decisions since November 21, except for 1. Each time, I set a full position with a stop loss of 0.5% to 1%, resulting in a total profit of about 5 percentage points.
1 - Sell ATM Put
2 - Sell ETH OTM on both ends, buy ATM + buy more Call options
3 - Buy spot during 15-minute level volatility compression
4 - Set up ETH 3100～3200 ladder spreads at a 1:3 ratio; despite a sudden surge causing chaos, DDH still lost some money
5 - Buy spot during 1-hour level volatility compression

Sold the Apocalypse ATM 83000 PUT, approximately equivalent to 82000 for a transaction.

The plan made earlier was to buy around 90,000, expecting at least a rebound, because I've decided to buy some, so I'm selling the 90,000 Put for 3D and waiting slowly.

Why does GPT Pro also experience rate limiting when using Codex? Is this a new restriction? I haven't encountered it before.

Yesterday I thought of an interesting trading strategy:
Long BTCDOMUSDT + Long ETH/BTC ( adjust the ratio )
=
What is the short-selling ratio of altcoins?
=
Shorting the altcoin index ( long position version )?

I have repeated countless times that if there is a liquidation, at least there will be a Rebound.
This morning, I sold a Monday expiring ETH 3500 Put. If it drops, I will buy, and if it doesn't drop, I will continue to sell the Friday 3400～3500 Put.
The probability of this plan outperforming spot bottoming is high without overselling.
1. If there is a fluctuation, the spot may not receive goods, and it is unclear how much it will rebound?
2. Continue to plummet, equivalent to buying spot at 3500, even if the extreme market drops to 3100, it should rebound, at least there is a chance to reb

I have mentioned countless times that if there is a liquidation, at least there will be a Rebound.
Sell ETH 3500 Put expiring on Monday in the morning. If it drops, I'll buy; if it doesn't drop, I'll continue to sell Friday 3400～3500 Put.
This plan is likely better than spot bottoming.
1. Next, there will be fluctuations, and the spot market may not receive goods, unsure how much it will rebound.
2. Continue to plummet, equivalent to buying spot at 3500. Even in extreme market conditions, if it falls to 3100, there should be a Rebound. At 3500, there is at least a chance to rebound

Let me review a bit and share some experiences on how to reduce losses.
1. Market risk, I will avoid the possibility of forming a LL pattern on 4H or higher. As long as it appears, I will decisively reduce leverage and cut positions. Even if it doesn't drop afterwards, it doesn't matter. I can accept missing out, but I do not accept being heavily invested at the starting point of a potential downward trend.
2. There is an arbitrage exposure during the drop, because I worry about extreme market conditions ADL, so I turned on all alarm notifications and kept the arbitrage positions very

I'll briefly review and share some experiences to avoid losing money.
1. Market risk, I will avoid potential LL formations on 4H and above. As soon as they appear, I will decisively reduce leverage and cut positions. Even if it doesn't drop afterwards, that's fine; I can accept missing out but I won't accept being heavily invested at the starting point of a potential downward trend.
2. There is an arbitrage exposure during the downturn, because I worry about extreme market conditions ADL, I have opened all alert notifications and kept the arbitrage positions very small. Woke up

Woken up by alarms in the middle of the night, I was flustered trying to balance the arbitrage legs, and in the end, I neither made a profit nor incurred a loss, so be it🥲🥲🥲

A very practical tip discovered during AI programming practice is to let him:
What is the core issue in a nutshell, and how should it be resolved?
This allows you to set aside all the details and understand what AI intends to do, making decisions purely based on logic to see if it meets your needs.
This kind of highly concise and to-the-point summary can only be achieved through communication with AI. When talking to a person, the other party will likely become very tired and won't be able to sustain it for long.

The changes in AI large models are really fast. Due to Claude Code's limitations, it has already become an assistant to Codex for me. I discuss solutions with CodeX, then ask Claude Code about the parts I don't understand, align the code, write documentation, submit to git, and let Claude Code handle other tedious tasks. The core work is done by Codex.

The changes in AI large models are really fast. Due to Claude Code's decline in intelligence, it has already become an assistant to Codex for me. 🤣 I discuss solutions with Codex, and then ask Claude Code about the parts I don't understand. Codex handles the core work, while Claude Code takes care of code alignment, writing documentation, submitting to git, and other tedious tasks.

Reviewing my journey in the secondary market:
1. Looking for indicators, technical analysis methods, and then going all in for a one-time solution. Holding onto losses, but taking profits when they arise.
2. Can control FOMO, lacks the courage to place bets. Neither losing money nor making money.
3. Recognize that the market is unpredictable and the importance of risk management. Understand your own risk tolerance and let go of unrealistic fantasies about trading.
4. Ultimately, everything must come down to execution. Find ways to simplify, adjust positions, and make the plan easy to implement