Hey @ClaudeDevs why do I have to vibe code my own implementation of copy-on-write for worktrees?
Should be native (and default if running APFS)
5
10
401
hey @mert has @zodl_app been removed from the @solanamobile seeker app store?
I can't find it, or update it any more?
2
11
406
Need an adversarial AI security council to inspect and verify all the things my bots want to run
Still doesn鈥檛 feel safe to let them go free, but I am a choke point for them
4
11
416
Robinhood is giving partial fills on market orders for $zec
1
8
887
wtf @Rabby_io
Unlocking the wallet automatically signs the last open pending transaction?
How on earth?
What happens in those ten minutes:
- As the game starts, the page queues a signature request permitted to drain funds.
- A pop-under bug in chrome hides the Confirmation dialogue, so the user sees nothing.
- After 10 minutes, the wallet auto-locks with the request is still pending, this does not cancel the request.
When you type your password to unlock, that password is consumed as approval for the request instead. It is signed and returned to the page without the user ever seeing or confirming anything.
5
9
1,037
We resolved the vulnerability upon discovery and released an update on August 11. Please ensure your Rabby extension is up to date. The mobile app is unaffected.
The conditions required to trigger this vulnerability are extremely limited:
1/ The wallet must be connected to a malicious website.
2/ The user must have manually set their auto-lock timer to specifically 10 minutes (all other timer settings are completely unaffected).
No exploits have been detected in the wild.
2
192
So is it even possible to sell a BIP110 bitcoin?
Or is the chain just locked up because they don鈥檛 have the hash to produce another block
Tick tock, next blo鈥evermind
4
7
557
Leaving fable 5 running a /goal overnight used to be like like leaving a 10X long open. Rarely you wake up pleasantly surprised, but almost always you wake up poorer
Running Deepseek like buying spot - it's never going to excite me much, but it's also not setting money on fire
2
10
274
You should know that when using the @AwakenTax API - it's much more powerful than just tax advice -
3
8
280
wholistic retweeted
there doesn't have to be an end goal
you can just make useful things for people who want to use them, and live a happy life
33
223
2,038
274,113
This weeks AI meta, we know all the models make mistakes, but if they are cheap enough and good enough you can just have them adversarially audit themselves in massive parallel agentic armies until the results converge on good enough
And with deepseek this costs a few cents
226
I am willing to pay 10X more for even a marginal improvement in intelligence
I will still use the marginally worse intelligence prolifically, but if something is actually important, or getting audited for production, I will pay whatever the best model is charging.
3
324
It鈥檚 actually unbelievable how unreliable osx spotlight search is at finding my applications
I can type the first 5 characters of the name, and it can鈥檛 find it in a list of ~200 items
2
3
465
requiring some user generated entropy should be the norm any time randomness is actually required, and I should be able to add as many bits as I like (within reason)
far as I know @Ledger doesn't allow you to add your own entropy to seed phrase generation
2
4
327
You are auditing a new audit procedure. Audit pack is prepared, audit pack was prepared by same procedure you are auditing. So extend the audit beyond the scope of the audit pack itself to capture any meta issues that might be self-referential, and obscure a deeper discovery.
3
4
252
