The more useful my agents become, the less I want one provider to own the context they run on.
Models will keep changing. The context around the work is what compounds.
Open-weight models are moving fast enough that I would not build a business workflow assuming one provider stays ahead forever. Optionality matters more once your context becomes valuable.
📢Meet Qwen3.8-Max — our most capable model to date.
Next week, the open weights of Qwen3.8-Max will be released, and Qwen3.8-27B is also going open-weights to meet you all!🎉
Qwen3.8-Max, a new bar for coding and cowork at 2.4T parameters:
- Autonomous coding: 10+ days of self-evolving development, from empty folder to production without hand-holding, complete project trace in the GitHub:https://t.co/iVHZWQoeSo
- Real work, real results: Production-quality deliverables across hundreds of professions.
- Long-horizon mastery: System-level autonomous planning with closed-loop adaptive learning, driving 500+ turns of chip design optimization and 365 days of e-commerce strategy.
- Native multimodal intelligence: Vision isn't just input — it's a continuous feedback loop for planning, execution, and self-correction.
💰Pricing:
Input: $2.0 / M tokens
Output: $6.0 / M tokens
Implicit Caching: $0.25 / M tokens
Start building with Qwen3.8-Max! 🚀
📖 Blog: https://t.co/iwjmQxLBof
✅ Qwen Studio: https://t.co/4V2pFvDovG
⚡ API: https://t.co/gAGqaLQGbN
Last week, 12 people voted on the direction of the dollar. The decision was made 9 to 3.
Bitcoin has spent months failing to settle a proposed rule change.
To some that may look like dysfunction. But if you really understand Bitcoin, you realize that is the whole point.
The upgrade nobody can approve:
https://t.co/6zHfWfHBaO
@kinetic_finance Agreed was one of the things I was most surprised about reading during this whole event.
Risk needs to be spread across multiple wallets and vendors.
The goal of Bitcoin isn’t to make self-custody a test of technical competence.
The goal is to give people genuine ownership of their money.
If the path to that ownership is too complicated, most people will hand it back to a custodian. That is the problem we need to solve.
AI agents could become one of the biggest productivity shifts of our lifetime.
But it cannot become the Wild West, where people hand over their data, inbox, money, and systems without understanding what can happen next.
The winners will make agency powerful and bounded.
The recent self-custody failures are awful, and the Bitcoin industry has to take responsibility for that.
But if our answer is that billions of people should roll dice, manage complex backups, and hope they never make one mistake, self-custody will remain niche.
We have to make it radically simpler without giving up sovereignty.
This is a wake-up call.
Bitcoin adoption will not come easy, and self-custody cannot stop at buying a device.
We need better defaults, better education, and setups that can withstand failure.
Learn from it. Improve it. Keep building.
The lesson from COLDCARD is not that open-source code is too dangerous in the age of AI.
It is that we should assume every security-critical codebase will eventually be examined by a tireless adversary.
Design for discovery. Then make sure one bug cannot become total loss.
Everyone’s talking about moving from self-custody to ETFs after the recent Coldcard news.
But almost nobody is talking about the tax implications.
For most people, you can’t simply “move” your Bitcoin into an ETF. You first have to sell your BTC, potentially triggering a significant capital gains tax bill, before buying ETF shares
👋I'm knowledgeable. Full disclosure, I lead the team that builds Bitkey. Gonna share some thoughts to help you make a decision and aim to be totally honest and transparent here. Hope you're OK with a long post so I can just share my full thoughts. This isn't canned, this is me just sharing my thoughts live.
What's great?
The 2-of-3 multi-sig setup is a superpower for a number of reasons. It helps protect against the type of things that happened in this incident. Our entropy is solid, but if it weren't on one of the keys, that's not alone enough to threaten your bitcoin. We generate entropy in 3 different environments (the phone, the hardware, and the server). This helps diversify, where a bug in one wouldn't be present in the others.
On top of this, 2-of-3 gives you flexibility where you can spend funds without Bitkey being involved. Even if @blocks disappeared (we have the Emergency Exit Kit for that). Block can be involved when you lose a key and need to recover.
You can also turn on a feature that lets you set a daily limit you'd like to spend on the go. With this on, our server will co-sign with your app key up to that limit daily, so you don't have to take your hardware if you'd like to spend while you travel or are out and about.
Our recovery system is unmatched. Lose your phone? Tap your wallet on your new phone and you're back in. Lose your Bitkey? Tap your new hardware, wait 7 days while we make sure nobody contests your recovery request, and you're in. Lose your phone and your Bitkey? If you've spent the 2 mins it takes to set up a recovery contact, they can help decrypt your app key from your cloud, without ever having access to it.
The 2 of 3 also allow us to provide an inheritance feature that is extremely easy to set up. It requires that your beneficiary have a Bitkey, but we give you a discount when you set it up and within minutes, your loved ones are protected.
We have a design without a seed phrase, which just means we take care of storing the seed for you in a redundant, resilient way. Because of this, onboarding is dead simple and takes less than 5 minutes and is easy enough for anyone of any age / experience level.
We're the only collaborative custody service that can't see any of your activity except one's we're involved in signing. This is because we invented Chaincode Delegation (https://t.co/jVGOV9GGeA).
We're fully FOSS. Our code is online and editable.
Where can we get better?
We need more utility features and we've already planned many of them. We've got support for transaction notes on the way, which is our first step towards ultimately supporting UTXO labeling and coin control. We're talking about how we can use this to support features like account separation.
We don't yet support lightning. Do we want to? Yes. Is it hard to do native lightning for mobile self custody? Yes. But there are lots of really interesting newish options that some type of lightning support much more straightforward and we will talk to the community about the right shape/tradeoffs for this when we make it past some of the other stuff I mentioned.
The lack of seed phrases is a tradeoff. We think it's the right one for most who self custody now and will in the future, but certainly not for all. If portability of your keys to other wallets is a crucial feature for you, Bitkey isn't right wallet. And that's OK. If portability of your funds is most important, then luckily we have Bitcoin transactions which allow you to move your funds from Bitkey to any other wallet.
So many of the things that make Bitkey great are because of the seedless decision. It reduces phishing capability, takes the burden of securing cryptographic materials off of customers, and in our opinion, greatly reduces the stress of self custody.
I'm not trying to do a sales pitch, so pardon if any of it sounds that way! There are other great products and services on the market as well. I do want people who have questions about our product to get the answers they're looking for. I'm really sorry you've gone through the roller coaster you've been on.
I disagree.
Self custody is not dead, but the standards across the industry need to rise.
If you truly believe in Bitcoin, the sovereignty aspect plays a huge role, which cannot be achieved without self custody.
Just because an institution holds your Bitcoin for you does not make it immune from hacks or confiscation.
Bitcoin is becoming the canary in the cybersecurity coal mine.
When software protects bearer money, bugs are not data breaches. They become immediate, irreversible losses.
AI will make finding flaws cheaper for attackers and defenders. The standard has to rise.
The people affected did not “fail at self-custody.”
Many followed the advice: buy a respected hardware wallet, keep the seed offline, and avoid exchanges.
The tool failed them. We can defend self-custody without blaming the people who trusted it.
Your Bitcoin custody setup should survive:
One device failing.
One company failing.
One backup being lost.
One person making a mistake.
If it cannot, it is not ready for a meaningful amount of bitcoin.
The right response to the COLDCARD disclosure is neither panic nor denial.
Check which device and firmware generated your seed. Follow the official guidance. Verify every step. Migrate carefully.
Rushing a wallet migration can create a second problem.