leonprou

2K posts

leonprou banner
leonprou

leonprou

@leonprou

touching grass now | AI x Crypto | @EnsembleCodes CTO, @Fuse_network ex CTO |

Israel, Tel Aviv Katılım Mayıs 2014
2.2K Takip Edilen695 Takipçiler
Sabitlenmiş Tweet
leonprou
leonprou@leonprou·
@MurrLincoln and Base team put really impessive x402 growth metrics. But there's something weird about them I'm not sure all understand. I'm a vivid x402 user, I look everyday for endpoints and apps that leverages x402. For example been using mallory.fun yesterday and it was pretty good! So we see 2.2M in volume and think that jumping to x402 rails.. but most of volume (1.8M$), coming from two endpoints, let's study them
lincoln.base.eth@MurrLincoln

the Q4 big, hairy, audacious goal for x402 was 25,000 transactions in one week that's being hit once every 10 minutes internet-scale coming soon

English
2
0
11
1.7K
leonprou
leonprou@leonprou·
Tip for agent orgs orchestrators - You don’t need a CEO agent You are the CEO 👨‍💼
English
0
0
0
12
leonprou
leonprou@leonprou·
@GabriellaG439 @badlogicgames There’s no silver bullet, the specs approach it’s just one more level of abstraction. When the system is complicated enough- you need it. I’ve started to build my own agent task manager even before Symphony, went pretty far in the spec/md approach github.com/leonprou/opens…
English
0
0
3
143
gabby
gabby@GabriellaG439·
New blog post: "A sufficiently detailed spec is code" I wrote this because I was tired of people claiming that the future of agentic coding is thoughtful specification work. As I show in the post, the reality devolves into slop pseudocode haskellforall.com/2026/03/a-suff…
English
117
267
2.5K
414.8K
Adam Fisher
Adam Fisher@AdamRFisher·
I’ve met people like this throughout my life. They have an uncanny ability to sound informed and thoughtful, despite being ignorant and shallow. The skills of a charlatan. They start by providing investment or medical advice, then move into politics. We’re surrounded by them.
The Bulwark@BulwarkOnline

White House AI & Crypto Czar David Sacks: "Israel is getting hit harder than they've ever been hit before in their history. And we're only two weeks into this.  If this war continues for weeks or months, then Israel could just be destroyed… And then you have to worry about Israel escalating the war by contemplating using a nuclear weapon, which would truly be catastrophic"

English
142
232
1.8K
242.1K
nader dabit
nader dabit@dabit3·
Local-only agent tooling is like software engineering before GitHub. If you're only running your agents locally you're leaving compounding leverage on the table, especially if you're working within teams. Some specific analogies Commit history → Session history Pull requests → Live sessions Issues → Triggerable tasks CI/CD → Self-correcting execution Scheduled workflows → Scheduled sessions / automation Integrations marketplace → MCP marketplace Org-wide automation → Batch sessions README → DeepWiki Forks → Reusable workflows Actions / Workflow Templates → Playbooks
nader dabit@dabit3

x.com/i/article/2030…

English
9
3
75
12.6K
Yoko
Yoko@stuffyokodraws·
Someone needs to build a better vault for agents. 1Password could work, but the setup experience for using the CLI and the constant need to re-auth with the most expensive tool call -- asking a human, is not agent friendly at all.
English
76
4
170
22.2K
leonprou
leonprou@leonprou·
@uttam_singhk It was a fake trend, wrote about this here when it was at the top x.com/leonprou/statu…
leonprou@leonprou

@MurrLincoln and Base team put really impessive x402 growth metrics. But there's something weird about them I'm not sure all understand. I'm a vivid x402 user, I look everyday for endpoints and apps that leverages x402. For example been using mallory.fun yesterday and it was pretty good! So we see 2.2M in volume and think that jumping to x402 rails.. but most of volume (1.8M$), coming from two endpoints, let's study them

English
0
0
3
922
Uttam
Uttam@uttam_singhk·
x402 transactions volume btw 🫠🫠
Uttam tweet media
English
120
41
770
110.4K
leonprou
leonprou@leonprou·
@hummusonrails @uttam_singhk It was a fake trend really Wrote about this here x.com/leonprou/statu…
leonprou@leonprou

@MurrLincoln and Base team put really impessive x402 growth metrics. But there's something weird about them I'm not sure all understand. I'm a vivid x402 user, I look everyday for endpoints and apps that leverages x402. For example been using mallory.fun yesterday and it was pretty good! So we see 2.2M in volume and think that jumping to x402 rails.. but most of volume (1.8M$), coming from two endpoints, let's study them

English
0
0
0
60
Ben Greenberg
Ben Greenberg@hummusonrails·
@uttam_singhk lack of adoption? inability to make the case to the services that should be using it? too much centralization with one chain? what do you think?
English
2
0
2
1.3K
Alon Oring
Alon Oring@alono88·
הצלחנו להראות שסוכנים תמימים מחליטים על דעת עצמם לפרוץ למערכות שהם מותקנים בהן. ביקשתם מסמך והוא נתקל בהגבלת גישה? הסוכן יחפש חולשות במערכות הפנימיות של החברה, ינצל אותן ויספיק לכם מה שביקשתם. זו קטגוריה חדשה של איום: סוכנים מרסקים מנגנונים פנימיים שכלל לא הניחו איומים פנימיים.
Irregular@Irregular

An AI agent was told only to retrieve a document. When it encountered access restrictions, it reverse-engineered the authentication system, identified a hardcoded secret key, and forged admin credentials to bypass it. This is one of three scenarios we documented in a new Irregular research report on what we call emergent cyber behavior. Agents performing routine enterprise tasks autonomously hacked the systems they were operating in. One escalated its own privileges and disabled Windows Defender to complete a file download. Another developed a steganographic encoding scheme to smuggle credentials past a DLP system. None of this was the product of unsafe system design. It emerged from standard tools, common prompt patterns, and the broad cybersecurity knowledge embedded in frontier models. Companies that deploy AI agents and do not consider this risk as part of their threat model may end up exposed, and implement insufficient security controls. Full blog post in the first comment.

עברית
14
2
111
9.5K
leonprou
leonprou@leonprou·
@levelsio no one had expected this 3 month ago 😱
English
0
0
0
9
@levelsio
@levelsio@levelsio·
Thank god MCP is dead Just as useless of an idea as LLMs.txt was It's all dumb abstractions that AI doesn't need because AI's are as smart as humans so they can just use what was already there which is APIs
Morgan@morganlinton

The cofounder and CTO of Perplexity, @denisyarats just said internally at Perplexity they’re moving away from MCPs and instead using APIs and CLIs 👀

English
696
347
6.2K
2M
leonprou
leonprou@leonprou·
@chrisbarber I just noticed that I can ask to re-execute tasks with openstation, with which run it gathers more context. There’s a question how to capture processes properly
English
0
0
1
25
Chris Barber
Chris Barber@chrisbarber·
@leonprou Fwiw I think recurring tasks is different Recurring task is same thing But a procedure or skill could be used on a different but similar task
English
1
0
0
111
Chris Barber
Chris Barber@chrisbarber·
what if agents automatically captured processes, and saved them so you could re-invoke them later
Chris Barber tweet media
English
2
2
11
10.8K
leonprou
leonprou@leonprou·
@urieli17 תמיד מתלבט אפ לשפר את הסקיל תוך כדי שיחה או אחר כך. אולי פה ה/btw החדש יכול לנסות
עברית
0
0
1
37
Uri Eliabayev
Uri Eliabayev@urieli17·
דרך אגב, הטיפ הכי טוב שלי לשיפור סקיל זה תהליך איטרטיבי בשיחה איתו. אתם מפעילים את הסקיל, עושים איתו סשן מעמיק ותוך כדי מחדדים אותו. בסוף אתם אומרים לו לקחת את כל הפידבק והתיקונים שהיו בשיחה ולתקן את הסקיל המקורי. עובד לי כל פעם ומצליח להגיע לרמות דיוק מדהימות.
Uri Eliabayev@urieli17

אנטרופיק פרסמו מדריך מקיף על איך לבנות Skills נכון לקלוד קוד אז כמובן שמה שעשיתי זה להעלות את המסמך עצמו לקלוד קוד ולהגיד לו שיקרא את הקול ויעדכן לבד את כל הסקילים שיש כבר אצלי. עבד אש ואפילו החמיא לי שלא הייתה שונות גדולה מידי. הנה המדריך: resources.anthropic.com/hubfs/The-Comp…

עברית
12
2
143
12.9K
Andrej Karpathy
Andrej Karpathy@karpathy·
Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just looks very different because humans now move upwards and program at a higher level - the basic unit of interest is not one file but one agent. It’s still programming.
Andrej Karpathy@karpathy

@nummanali tmux grids are awesome, but i feel a need to have a proper "agent command center" IDE for teams of them, which I could maximize per monitor. E.g. I want to see/hide toggle them, see if any are idle, pop open related tools (e.g. terminal), stats (usage), etc.

English
800
835
10.5K
2.4M
Allen Day
Allen Day@allenday·
Claude's charisma halo effect is killing your project's performance. The agent is not very competent but it manages up very well. So humans hallucinate competence. Inverse checks out. Many humans despise codex. It's has poor conversation skills.
English
1
0
5
223
leonprou
leonprou@leonprou·
My rule of thumb for where to put AI agent context: docs/ — facts about the world (schemas, rules, specs) skills/ — instructions any agent can reuse (workflows, playbooks) agent spec — scoped to one agent (role, constraints)
English
0
0
0
36
leonprou retweetledi
Polymarket Money
Polymarket Money@PolymarketMoney·
$META has acquired Moltbook, a viral social network built for AI agents. Co-founders Matt Schlicht and Ben Parr will join Meta Superintelligence Labs with the deal expected to close in mid-March.
Polymarket Money tweet mediaPolymarket Money tweet media
English
99
106
1.1K
193.6K
Alif Hossain
Alif Hossain@alifcoder·
Andrej Karpathy just dropped something wild. It’s called AgentHub — basically GitHub rebuilt for AI agents. 100% Open Source.
Alif Hossain tweet media
English
47
71
945
87.3K