For a macro-life. Aiken Core Maintainer. Cognitive Security Maxi
Higher Liquid
Joined October 2020
- Tweets1.3K
- Following219
- Followers2.8K
- Likes2.7K
Pinned Tweet
I asked chatgpt 5.4 to do the next step of the canonical ast I'm working on and it's considering a second job. What is up with this economy!?
#keep4o
📢MicroProofs retweeted
What the fuuuuuuuck it's real, this is 596 sessions just from clicking on articles
So I've decided to align myself better with AI, I must match their substrate. Time to add gigawatts of electricity to my daily regiment
AI Alignment appears slow because it is framed in the wrong direction. It's not AI alignment with humans. It's human alignment with AI.
📢MicroProofs retweeted
The Asset Recovery Tool Is Now Live
The official asset recovery tool is now available for users whose wallets were affected by the June 2026 security incident. It allows you to verify control of an affected wallet, review any assets eligible for recovery, and submit a recovery request to a new Cardano wallet that you control.
The recovery tool is available only at secondfi.io in a web browser. It is not completed through the SecondFi app. You do not need to have submitted a claims ticket before you start. Submitting a recovery request through the official tool begins the claims process for the affected wallet.
Recovery is separate from wallet migration. Migration moves supported assets that remain in a wallet to a new wallet. Recovery concerns eligible assets affected by the security incident.
How long recovery takes
There are two separate stages, each with its own timeframe:
-Proof generation, while you use the tool: up to approximately 15 minutes. This takes place on your own device, in your browser, and the browser window must stay open until it is complete.
-Claim processing, after you submit: up to five working days. To maintain security measures, submitted claims are processed in batches. No action is needed from you during this stage.
Before you start
Have the following ready:
The recovery phrase for the affected wallet. The tool supports 12, 15, and 24-word recovery phrases.
A newly created Cardano wallet that you control, and its receiving address. A hardware wallet can provide an additional layer of security because it keeps private keys offline.
A private, trusted computer with a current web browser. Do not use a public or shared computer.
A device and internet connection that meet our recommended specifications: at least 2 GB of available memory, a processor no more than five years old, and an internet connection of at least 50 Mbps. Recovery runs entirely in your browser, is demanding on both processing power and memory, and requires you to download a file of about 2 GB.
Do not use the affected wallet, an exchange deposit address, or any address you do not control as your destination.
How recovery works
Recovery is one continuous process. Begin by reading and accepting the Recovery Terms. The tool then takes you through four steps, set out below.
If you have more than one affected wallet, repeat the process for each wallet.
Step 1: Verify your affected wallet
Enter the recovery phrase for one affected wallet. Before you type it, confirm that you are on your own private, trusted computer, that nobody can see your screen, that no screen sharing or recording is running, and that the address bar shows secondfi.io.
Step 2: Review what is eligible
The tool checks the wallet's addresses against the incident records and shows an estimated breakdown of any assets eligible for recovery. Each affected address is treated as its own request, and only addresses with recoverable assets are submitted.
Step 3: Enter your destination address
Provide the receiving address of your new Cardano wallet and complete the required confirmations.
The destination address is bound into the proof and cannot be changed after you submit. Check every character carefully before you continue.
Step 4: Generate the proof and submit
The tool generates a zero-knowledge proof on your device for each eligible affected address, binding it to your destination address, and then submits your recovery request.
Proof generation can take up to approximately 15 minutes. How long it takes depends on your device's processing power and your internet connection, which is why we recommend the device specifications listed above. It may take longer on older or less powerful computers and for wallets with more affected addresses.
Keep the browser window open and do not leave or refresh the page until proof generation is complete. If it is interrupted, the tool will show a retry option.
After you submit
Once your request has been validated and submitted, it enters claim processing. Claims are processed in batches, so allow up to five working days for your request to be processed and for recovered assets to arrive in your destination wallet. No further action is required unless the tool or our official channels tell you otherwise.
Please remember: This is not instant, it will take up to five working days for your request to be processed.
You can check your status at any time using the status checker at secondfi.io by entering your new destination address, not the affected wallet address.
Why the tool asks for your recovery phrase
Your recovery phrase is used locally in your browser to generate the zero-knowledge proof. It is not stored or transmitted. Only the proof is submitted with your recovery request.
Browser and security tool warnings
Because the recovery tool uses your existing recovery phrase to prove ownership of your wallet, some anti-phishing tools, anti-scam tools and browser extensions may show a warning, or display a blank screen, when you open the tool or enter your recovery phrase. This alone does not mean the official tool is malicious.
Before you continue, always check that you are on the exact official domain, secondfi.io, with HTTPS and no extra words, characters or lookalike spellings. Always type the address yourself rather than following a link sent by direct message, email, advertisement or any other third party.
What can be recovered
Amounts are assessed against the incident snapshot and shown as an estimate. The final amount may differ.
For each eligible NFT that cannot be returned directly, a fixed amount of 30 ADA will be included in the ADA shown for the relevant affected address.
Rules that protect you
Type secondfi.io directly into your browser or from the official X account mentioned below. Never follow a link from an email, advertisement, direct message, or search result.
Enter your recovery phrase only yourself, and only in the official recovery flow after confirming that the address bar shows exactly secondfi.io. Make sure nobody can see your screen and that no screen sharing or recording is running.
Do not install any browser extension or application claiming to be required for recovery.
No fee is required. Anyone who asks you to pay a fee, send ADA or another asset, sign a transaction, or transfer assets to an address they provide is attempting to defraud you.
We will never contact you first to ask you to begin recovery.
Do not use the affected wallet again. Recovery does not make it safe to use.
Our official channels
X accounts: @secondfiapp and @secondfi_jp
Support portal: support.secondfi.io
📢MicroProofs retweeted
This post looks like the start of a VERY sophisticated and well-funded PR operation to get support for Democrats to regulate AI into oblivion. Let me show you how it works:
1.) This guy, with minimal followers and no previous account activity, goes to the Wall Street Journal which publishes an exclusive with quotes from him on his resignation 18 minutes BEFORE this post goes up. Planning was clearly done in advance.
2.) Within hours, it has tens of thousands of reposts and the account has 100k+ followers. The post is punchy, quotable, it almost seems professionally written. The first three accounts to quote tweet it all do so within 15 minutes of the initial posting. Remember, this account had basically zero engagement beforehand, so an organic reach explanation seems unlikely.
According to Grok those accounts are @_NathanCalvin (General Counsel at Encode AI), @peterwildeford (Head of Policy at the AI Policy Network), and @DKokotajlo (Head of the AI Futures Project), all of which are up-and-coming AI-Doomer policy advocacy nonprofits.
The AI Futures Project website says it is funded “primarily” by the Survival and Flourishing Fund, which says on its own website that it has advised Jaan Tallinn, Skype creator and one of the leading investors in Anthropic, to grant over $2.5 million to the AI Futures Project since 2024.
Encode AI says on its website that it is ALSO funded by the Survival and Flourishing Fund, which in turn says that it told Anthropic investor Jaan Tallinn to grant $516,000 to Encode AI in 2025.
And wouldn’t you know it, the Survival and Flourishing Fund ALSO says it told Jaan Tallinn to grant $2 million to the AI Policy Institute, the 501(c)(3) affiliate of the AI Policy Network, as well.
What are the odds that the first three quote tweets of Coxon’s post would all be major AI-restriction policy advocates funded generously by the same donor, who also happens to be one of the leading investors in, and a board member of, Anthropic, the company Coxon was resigning from? And all within 15 minutes of posting (two within ten)?
3.) Jacob Coxon doesn’t have much of a resume, but we do know that, in 2022, he got a $20,159 scholarship for the “long term future scholarship program” from the Good Ventures Foundation, one of the philanthropic vehicles of Dustin Moskovitz, a notorious AI-doomer who has spent tens if not hundreds of millions on policy advocacy to strictly regulate AI, while also being an Anthropic Investor himself.
It also just so happens that the 14th person to quote Coxon’s post was @MaxNadeau_ (27 minutes after posting) who is the program officer for the Technical AI Safety team at Coefficient Giving, another of Moskovitz’s philanthropic spending vehicles. Max is not a frequent poster, his last posts before quoting Coxon were before Labor Day, but he was remarkably quick off the mark for this one.
4.) Basically every major Democrat politician and candidate has suddenly glommed on to this post, and conveniently, as the people cry out foe answers, Bernie Sanders already has a bill written to “ban super intelligence” and regulate AI into oblivion, and will be releasing later this week. The bill, among many other things, will create “a new cabinet-level federal agency to safeguard the public from the dangers of artificial intelligence” that will be “advised by an Artificial Intelligence Advisory Board comprised of experts on artificial intelligence.” Do you think, perhaps, Anthropic and its many investors who fund AI policy advocacy might have interest in getting to place a pet “expert” on the board of an entity that dictates what AI is and isn’t allowed to do? And isn’t it fortuitous that this whistleblower came forward with his oh-so scary stories so close in proximity to the release of the most radical piece of AI legislation ever introduced?
Yeah it's the curse of either lower thinking more involvement and more iterations or higher thinking plus compounding slop creation
Higher reasoning often results in higher task accuracy (higher chance that it will implement something that meets your defined success criteria).
But this doesn’t come for free, it comes with more tests, more task diligence, more checkpoint tests and defensive code.
The above ultimately can be summarized as “more slop”. More code bloat, more superfluous validation, more functionally useless smoke tests.
📢MicroProofs retweeted
🤖Agent harness hackathon. One long day.
Cross Exam, with @tomasdm_eth and @MicroProofs:
Rehearsal for the move you can’t take back.
$840 claimed. $96,310 on the replica. 611 duplicates.
❌Denied, then it fixes itself.
✅Every guardrail passed.
They score the guess. None of them execute.
You can probably guess the zip code. 🌉
📢MicroProofs retweeted
We can finally talk about it:
We found a way to extract hidden reasoning of frontier models using a vulnerability in the APIs of every frontier AI company.
We verified that our reasoning token count matches billed API thinking tokens 1:1 for most of the prompts we queried.
Got an opencode go subscription usage unable to get ahold of support after emailing
@opencode
Daily reminder don't get gaslighted by LLMs (even Fable, Sol, etc.)
More than 1-2 issues with some standard tools is indicative that the LLM ignored some dead simple principle and decided instead of backtracking to monkey patch in some wild shit.
The hardest but most important lesson is to understand at what point to stop rabbit holing and actually go down a different path, make sure you didn't miss a fundamental approach.
Note in my case this is just me messing around with a spike, not intended to be used anywhere just teach me some ideas
Majority of reboots, spin offs, late sequels recently
"I understood how important it was to get this right for them"
"I'm a fan too"
Proceeds to fuck it up for the fans immediately after
Classic
📢MicroProofs retweeted
I assure you nothing has changed. You may not have to manually type like before. You may even be moving faster than before. But the fundamental nature of the job of a SWE has not changed at all. If you think otherwise you probably misunderstood what the job was this entire time.
