Motorola Atrix dream from 2011 is finally delivered!
Henge 0.2 is out, and it now runs macOS 27 Golden Gate on an iPhone.
On a 12GB iPhone, the Mac gets 10GB of RAM. Sequoia still works too, and you can now give the Mac anywhere from 1GB to 16GB.
Also new in 0.2:
• USB-C external display support
• 60Hz or 120Hz ProMotion
• • Fixed colors, with P3 color matching on iPad
• iPad gets Retina and Performance display modes, better Magic Keyboard scrolling and a pointer size setting
• JIT can be turned on right from the Install tab
• You can sign it with a free personal developer account
Same requirements as 0.1: iOS 17 or later, A14 or M1 and newer, about 43GB free, and JIT.
github.com/Leviidev/Henge…
0
0
3
130
frol.near retweeted
I just tried to buy some Pearl coins via @NEARProtocol and holy shit that was the smoothest on-chain experience I ever had
It takes few clicks from having no NEAR wallet / receiving a gift via one of your exiting crypto wallets, to buying such an "inaccessible" asset like PRL
(note: we are long NEAR and PRL, but its my first time trying it out personally smh)
It takes few clicks from having no NEAR wallet / receiving a gift via one of your exiting crypto wallets, to buying such an "inaccessible" asset like PRL
(note: we are long NEAR and PRL, but its my first time trying it out personally smh)
25
23
286
36,001
frol.near retweeted
Live Cybersecurity Webinar!
Mainstream AI won’t help you find vulnerabilities in your code.
Our uncensored models will, privately, inside secure enclaves.
@zmanian shows how to hack your own code before attackers do.
🗓️ Nov 12
⏰ 10am PT
Register: luma.com/f0ixwvkx
Mainstream AI won’t help you find vulnerabilities in your code.
Our uncensored models will, privately, inside secure enclaves.
@zmanian shows how to hack your own code before attackers do.
🗓️ Nov 12
⏰ 10am PT
Register: luma.com/f0ixwvkx
5
7
65
8,256
Discord is dropping Electron and rewriting the whole app in Rust.
NEAR nodes - Rust. NEAR smart contracts - Rust. Trezu backend - Rust.
Rust is not a hype anymore, it is the default for software you actually care about. Welcome to the club, Discord.
NEAR nodes - Rust. NEAR smart contracts - Rust. Trezu backend - Rust.
Rust is not a hype anymore, it is the default for software you actually care about. Welcome to the club, Discord.
We've been early adopters of Rust since 2018 and a lot of our infra, voice tech, and even some of the current client is already Rust, the holy grail was getting the whole client there. That means Electron goes away, and so does React Native, mobile keeps native UI but shares app ust. We hope to get it into testing with users soon™. (And it's not Tauri, it's GPUI, so good guess everyone.)
Alright, back to staring at the WoW Forever login screen.
3
1
20
2,767
frol.near retweeted
Releasing Delta League: edit LLM weights for benchmarking and glory! Join at deltaleague(.)ai
Earlier this week, the idea of editing weights directly with an LLM got a lot of attention —
but it was shunted by limitations in the demonstrated approach.
Delta League is an open e to submit deltas to model weights and get them scored on a hidden test set. Auto-research and hillclimb to make smaller models more performant.
The broader vision for Delta League is creating a platform for open AI research. Beyond just editing weights, this approach offers verifiable evaluation and an open collaborative space for everyone to continue improving Pareto of open source models.
In the coming days, we will extend this platform to support editing model architectures beyond just weights and custom evaluation sets to understand how models perform in specific domains and for businesses.
Earlier this week, the idea of editing weights directly with an LLM got a lot of attention —
but it was shunted by limitations in the demonstrated approach.
Delta League is an open e to submit deltas to model weights and get them scored on a hidden test set. Auto-research and hillclimb to make smaller models more performant.
The broader vision for Delta League is creating a platform for open AI research. Beyond just editing weights, this approach offers verifiable evaluation and an open collaborative space for everyone to continue improving Pareto of open source models.
In the coming days, we will extend this platform to support editing model architectures beyond just weights and custom evaluation sets to understand how models perform in specific domains and for businesses.
56
40
293
41,565
In May I wrote “the interview scams are real”. Now it is the root of the $6.3B Binance hack.
Tampered Homebrew → fake job interview → “coding assignment” → internal AI tool → container deploy permission → hot wallets.
4 months, 5 steps. Any single one of them should NOT be enough to move the funds.
Tampered Homebrew → fake job interview → “coding assignment” → internal AI tool → container deploy permission → hot wallets.
4 months, 5 steps. Any single one of them should NOT be enough to move the funds.
3
1
14
1,526
The interview scams are real. Do not follow random links without verification (keep reading to learn what to pay attention to).
The story: We are running @NEARN_IO gigs board for NEAR ecosystem, and we recently got listings that looked legit, but also a bit AI-generated (not a bsite was also a bit suspicious, but no clear evidence of scam.
So, I decided to become the honeypot and applied for the role, and they maintained a reasonable level of communication with human-like delays with this scripted first response:
> I’ll tell you a little more about myself so you know who you’re talking to. I’m the Community Marketing Manager at https://<fake-website>/team. Right now, our HR manager has left the team, and I’m filling in for him.
I had to go further and got to the point of the interview link and they sent me a fake "Microsoft Teams" link, which looked like "microsoft. teams02eu .com" (please, don't follow the link!). When I said that I cannot use Mircosoft Teams, and offered my link, they refused to use it, and suggested a new link (claimed it to be Slack): "slack. host02space .com".
Needless to say that I immediately took down all the NEARN posts.
I have not followed the links, so I don't know what exactly would have happened, but clearly nothing good.
Whois service says that both domains are registered to the same company around the same time:
who.is/whois/teams02e…
who.is/whois/host02sp…
Stay safe!
0
0
1
33
This is exactly the threat you must be concerned about with running your agents. All it takes a single malicious skil, compromised docs/README file, or npm package with supply chain attack, and your machine is now part of the malicious network spreading the virus through the packages and projects you maintain.
DO NOT RUN your agents on your local machine without at least sandboxing them in a virtual machine. Ideally, run them on a separate physical computer.
The environment where agent can execute local commands must never have your personal accounts credentials! Create new GitHub account(s) and email(s) per work scope and grant read-only access to any data you want it to access (e.g. forward emails that you need it to access, or grant read-only access to your private repository and let it fork privately so agent can only submit PRs and never can it push unnoticeable to your production repo)
DO NOT RUN your agents on your local machine without at least sandboxing them in a virtual machine. Ideally, run them on a separate physical computer.
The environment where agent can execute local commands must never have your personal accounts credentials! Create new GitHub account(s) and email(s) per work scope and grant read-only access to any data you want it to access (e.g. forward emails that you need it to access, or grant read-only access to your private repository and let it fork privately so agent can only submit PRs and never can it push unnoticeable to your production repo)
The issue isn't you pushing bad code. The issue is your agent getting compromised and pushing code to production without you even knowing about it.
2
0
11
343
Follow this guy
Remove your main @github account from your computer. Now.
Give your AI agent a separate GitHub account with minimal permissions per repo.
Let the agent create PRs, then use your main account to review and merge them.
Obvious advice, but essential in the age of AI agents.
3
1
17
880
frol.near retweeted
Remove your main @github account from your computer. Now.
Give your AI agent a separate GitHub account with minimal permissions per repo.
Let the agent create PRs, then use your main account to review and merge them.
Obvious advice, but essential in the age of AI agents.
Give your AI agent a separate GitHub account with minimal permissions per repo.
Let the agent create PRs, then use your main account to review and merge them.
Obvious advice, but essential in the age of AI agents.
250
119
2,572
376,360
frol.near retweeted
When doing security research it's important to run it securely to avoid impact on your machine.
Use NEAR AI RED Sandbox to leverage uncensored models and isolated microVM for all tool calls.
Use NEAR AI RED Sandbox to leverage uncensored models and isolated microVM for all tool calls.
Had a little fun with terminal ui and launched a thing...
we gave NEAR AI's new cybersecurity models a burner VM. No network, no keys, no regrets. Feed it the binary you're scared of, watch it work, then torch the box.
Enjoy 🤘
$ rts launch
github.com/nearai/nearai-…
49
33
292
102,332
frol.near retweeted
Gemini is NEAR
Near has landed on Gemini
near:native is now available to trade, with deposits and withdrawals enabled
→ exchange.gemini.com/buy/NEAR
7
22
220
5,649
frol.near retweeted
We are in a wild time for building software, so I thought it would be helpful to lay out my rules for building in the age of AI. x.com/i/article/2114…
1
3
13
787
frol.near retweeted
I'm working on a new video answering questions about privacy and AI in products and general use.
I have scrapped some common questions, but if you have ones you want answered, ask below 👇
I have scrapped some common questions, but if you have ones you want answered, ask below 👇
0
3
9
451
