Catalin

479 posts

Catalin banner
Catalin

Catalin

@dcatalin98

Writing immutable bugs on $ETH & $SOL 💀 | Solidity • Rust • TypeScript | Mass adoption loading... ██░░░░░░░░ 10%

Bucureşti, România Katılım Eylül 2016
515 Takip Edilen126 Takipçiler
Sabitlenmiş Tweet
Catalin
Catalin@dcatalin98·
shipped mpc-agent-wallet v0.1.0 finally, a way for AI agents to sign transactions without holding private keys • 2-of-3 threshold MPC (DKLs23 protocol) • EVM + Solana • TypeScript, Python, Rust SDKs • ElizaOS + LangChain plugins MIT licensed, open source github.com/Kazopl/mpc-age…
English
3
1
18
1.6K
Catalin retweetledi
Safe{Labs}
Safe{Labs}@SafeLabs_·
Approve transactions from anywhere. Safe{Wallet} mobile: full transaction decoding, Tenderly simulation, Blockaid scanning, real-time notifications. Audited by Certora. The multisig experience that doesn't compromise when you're not at your desk. safe.global/mobile?utm_sou…
English
12
2
20
1.2K
Catalin
Catalin@dcatalin98·
@PaulFrambot how does rate discovery work without utilization curves tho? in a pool you at least have a pricing anchor. offer based matching where both sides post different terms... feels like MEV would have a field day
English
0
0
1
213
Paul Frambot 🦋
Paul Frambot 🦋@PaulFrambot·
Morpho Midnight is by far the hardest and most ambitious thing I have ever built in my life. We started more than 1.5 years ago and did so many iterations, and bootcamps with the protocol team. It has been years since DeFi saw a true 0-to-1 innovation in protocol design. I believe Midnight is not only the next one, but the largest fundamental unlock DeFi has ever seen, especially if you are an institution. Can't wait to share more when audits finish.
English
18
14
295
28.4K
Catalin
Catalin@dcatalin98·
@jwsaml been using cursor plan mode alongside both. cursor for daily work because the harness + UI is better. ultra plan for big multi file architecture. codex when i want minimal hand holding. no clear winner, different tools for different scopes
English
0
0
1
118
Jesse Samuel
Jesse Samuel@jwsaml·
Which plan mode is better? 1. Claude Code ultra plan 2. Codex plan mode
Deutsch
36
0
38
3.2K
Catalin
Catalin@dcatalin98·
@yacineMTB step 4 of The Algorithm only works after step 2. most teams try to speed up processes that shouldn't exist. then they slap AI on top and wonder where the 10x went
English
0
0
1
146
kache
kache@yacineMTB·
I used to be worried about the competition catching up to my team. Due to AI, the speed of programming, development is going to increase, and put them asymptotically right by us. But instead, the gap is increasing. Our speed is increasing too. We're getting further ahead
English
26
6
413
16.3K
Catalin
Catalin@dcatalin98·
@Alfred_Lin most orgs actively punish this behavior. you surface 30 broken things in one day, congrats, you now own 30 problems nobody wanted to look at. the missile only survives when leadership actually wants to hear it, not just says they do
English
4
0
36
5.7K
Catalin
Catalin@dcatalin98·
@icanvardar how long can they keep pricing below cost to lock enterprises in though. cloud wars worked because margins eventually flipped. with inference demand outpacing optimization, i'm not sure that flip ever comes
English
0
0
3
32
Can Vardar
Can Vardar@icanvardar·
i don’t think current llm pricing is sustainable long term unless compute costs drop hard either margins are way thinner than people think, or prices are gonna have to go way up to make the economics work
English
51
1
63
5.1K
Catalin
Catalin@dcatalin98·
@auren i'd push back on 'AI skills' as a category. the people i've seen get the most out of these tools already knew how to decompose problems before AI showed up
English
0
0
1
123
Auren Hoffman
Auren Hoffman@auren·
what are best practices on sharing AI skills in an organization?
English
50
1
42
8.5K
Catalin
Catalin@dcatalin98·
@yacineMTB it's mutually assured deployment at this point. neither company has a release schedule, just google alerts for each other
English
0
0
1
71
kache
kache@yacineMTB·
anthropic 5 milliseconds after openai releases their next model: "hey actually guys mythos is available in GA now haha please use it"
kache tweet media
English
82
87
4K
161.7K
Catalin
Catalin@dcatalin98·
@nikitabier you hired Aave's former CPO (who built a self custody wallet), announced in-app trading, x money launches this month. only "maybe" left is custodial vs self custodial
English
0
0
0
14
Nikita Bier
Nikita Bier@nikitabier·
Crypto has had a rough year. Maybe we should launch something to fix it.
English
12.6K
5.1K
43.3K
6.5M
Catalin
Catalin@dcatalin98·
@awsdevelopers at this point i'm convinced aws randomizes your region on login just to keep you humble
GIF
English
0
0
2
216
AWS Developers
AWS Developers@awsdevelopers·
when you log into the console and you're in the wrong region
AWS Developers tweet media
English
7
40
526
20.3K
Catalin
Catalin@dcatalin98·
@Scroll_ZKP dissolving a 12 member security council first and promising to "work with stakeholders to find a new structure" later isn't a transition. it's a gap. you design the replacement before you tear down what exists, not after
English
0
0
3
562
Scroll
Scroll@Scroll_ZKP·
Today we'd like to announce an important proposal regarding Scroll's protocol and governance operations. We're proposing to dissolve the Security Council and transition protocol admin control to a Scroll Admin multisig. We expect the transition to take place over the next ten days. The council's cost relative to its actual usage over the past quarters no longer justified continuation. These resources are better directed toward building products and driving growth. We've carefully weighed the security considerations that come with this change and are confident this is the most responsible allocation of resources at this stage. Security councils play an important role in the maturity of any protocol and that's why we will be working with key stakeholders to identify a new security council structure better adapted to current market conditions. Our infrastructure is mature, thoroughly audited, and actively maintained by the engineering team that built it. Scroll will continue operating under the highest security standards to safeguard users.
English
121
19
195
254.6K
Catalin
Catalin@dcatalin98·
@pmitu say no to 99 good ideas and bet everything on one. ai generates options, founders kill them
English
0
0
3
18
Paul Mit
Paul Mit@pmitu·
What can a founder do that AI can't?
English
299
0
184
21K
Catalin
Catalin@dcatalin98·
@PeterDiamandis no retraining, no calibration, no fine tuning. every model currently deployed can just... use this? the 6x memory reduction being a pure drop in might be bigger news than the numbers themselves
English
0
0
1
43
Peter H. Diamandis, MD
Peter H. Diamandis, MD@PeterDiamandis·
Google DeepMind winning again -- just dropped TurboQuant, reducing AI memory usage by 6x and attention computation by 8x without any accuracy loss. Cheaper AI = more AI experiments = faster iteration = better models = cheaper AI. It's a compounding loop... And we're entering the acceleration phase.
English
79
99
887
38.2K
Catalin
Catalin@dcatalin98·
@cryptodotnews european banks discovering they need stablecoin partners the same quarter MiCA enforcement hits. the urgency is touching
English
0
0
2
73
crypto.news
crypto.news@cryptodotnews·
JUST IN: Major European banks and corporates now actively selecting stablecoin deployment partners
crypto.news tweet mediacrypto.news tweet media
English
18
17
132
5.3K
Catalin
Catalin@dcatalin98·
buried 3000 words deep: aave labs will invest in agentic AI protocols building proper agent infrastructure early (permissions, spending constraints, not just API wrappers) are going to own the next wave of DeFi volume. this deserves way more attention than a bullet point
Stani@StaniKulechov

Aave Will Win, the most important proposal in Aave's history just passed with a landslide. Here's the master plan going forward: General Direction - Aave becomes fully token-centric: one asset, one model: $AAVE - To date, protocol revenue per AIP-1 has accumulated to the Aave DAO: $140M in 2025, with 2026 on track to match that despite the market downturn being limited to protocol-only revenue - The AWW proposal introduces a new revenue stream: application and product revenue generated outside the Aave Protocol, now directed to the DAO as additive revenue - This covers Aave Pro, Aave.com, Aave App, Horizon (RWAs), and Aave Kit, all flowing back to the DAO treasury - Swaps on Aave.com and Aave Pro are already generating $10–20M in new revenue on top of existing protocol revenue - Aave V4's reinvestment feature ensures that float capital in pools generates yield, creating additional revenue streams, similar to how Aave V4 Spokes open up new revenue opportunities - AWW gives Aave exposure to the full vertical stack. Owning that stack is increasingly critical in a competitive landscape where protocols get commoditized - AWW also establishes a community-protected vehicle to independently govern Aave's brand assets and IP on behalf of token holders - Aave Labs commits to working exclusively on Aave-related products, fully locked in - If you own $AAVE, you own not just the economic rights of the protocol, but the brand, the users, and the integrations In other words: everything belongs to one asset, the $AAVE token - We believe tokens are the greatest opportunity of our time to build collectively governed protocols, but a single, unified vision is essential for execution - That vision is provided by Aave Labs, working alongside all Aave service providers to grow Aave from a $40B protocol to $1T and beyond Product Layer & Distribution - Aave App will onboard millions of users with a simple, fintech-like experience while ensuring users retain full control over their funds, backed by $1M account protection per user. A card will also launch later, generating additional fees for the Aave treasury - Aave Pro will be the premier destination for power users: sophisticated features, simple on-ramping, and the best of DeFi in one place - Aave Labs has the best designers and design engineers in the space, committed to delivering a high-fidelity experience for every user - Aave Kit will provide SOC2-compliant, enterprise-grade integration for fintechs and partners - Horizon will expand with Aave V4 support and more flexible asset onboarding to scale RWAs on Aave - New Aave V4 Spokes will unlock additional collateral and address the demand side of DeFi liquidity - Together, these products aim to bring DeFi to everyone and position Aave as the base credit and repo market for the entire $400T+ TradFi asset base Engineering & Tech - Aave Labs has the best engineers in DeFi. We built V1 through V3, GHO, and most recently V4, and this is just the surface of what we're building next - Aave V4 paves the way for next-generation lending, and Aave V3 will remain fully supported and maintained by Aave Labs for years to come - We are security-first. Smart contract security, application security, and ICT security are non-negotiable, and our recent SOC2 compliance reflects that. Institutions expect it, and we deliver it - We will invest in agentic AI, opening up new opportunities for developers building with Aave Marketing - Aave has historically led crypto in brand, events, content, and partnership marketing. We're doubling down on our brand recognition and the strong foundation we've built - Going forward, we'll expand into new audiences and channels to bring Aave mainstream, building net-new, stickier userbases among people who are new to DeFi Growth - Aave will deepen relationships across the DeFi ecosystem and build new bridges with fintechs, banks and asset managers - At its best, Aave isn't a bank. It's a financial network that any fintech, bank or an asset managers can plug into, and providing the best integration tools will be key BD efforts will rely on tight collaboration between service providers such as Token Logic and our partner networks - We honor our long-term partnerships and commitments, including @chainlink - We recognize the value Aave represents today and expect partners to approach us with the same respect Governance - We support a multi-contributor model for Aave and will continue to embrace it - We will oppose any vendor lock-ins or service providers that build products for themselves at the expense of token holders - We require full transparency from the SPs and no tolerance for relationship gating as all value needs to drive to Aave - Zero value leakage: everything built with Aave's funds must benefit Aave and be owned by Aave - SPs who align with these principles and commit to what's best for token holders will have our support on budgets, as long as they are reasonable - The DAO is taking a zero-bureaucracy approach: execution and skin-in-the-game are what matter. We are competing with some of the world's most efficient and well-funded organizations, and there is zero room for friction - Every SP will have real, measurable goals. Payments for posting governance proposals are over. We've already consolidated SPs to focus resources - Governance process improvements are coming in the months ahead: more efficiency, less politics Risk Management - We will continue to support a multi-layered risk management process encompassing both an economics risk layer and a technical risk assessment layer conducted by Aave Labs - Aave's risk management will include external risk managers such as Llama Risk and Token Logic for commercial and economic assessment. Aave Labs will also establish a permanent internal risk management function to coordinate and support external risk managers, making the overall system more resilient Building a Regulatory Moat - Aave Labs has spent years building a regulatory moat around Aave's products and deepening vertical integration - Aave is one of the only DeFi ecosystems operating at scale with regulated entities, including Push Virtual Assets Ireland, which is authorised as a CASP under MiCA, alongside a UK EMI-licensed entity - We are actively pursuing additional licenses globally to enable seamless, 1:1 fiat-to-Aave onboarding for mainstream users, a prerequisite for mass adoption - We go where the bar is high Policy - Aave Labs' policy team is world-class. We've participated in every major policy consultation over the years and will continue to fight for DeFi, protecting it from harmful regulation and ensuring legal certainty for users and integrators - The next few years will be pivotal for DeFi policy. We are fully committed Our Principles - Security-first above all else. This is non-negotiable Everything we build is truly DeFi, with self-custodial access at its core - Innovation-driven, we will move the space forward by innovating and building something new - For DeFi to scale, we need new audiences. That means growing the pie by building better experiences and infrastructure for users to access DeFi - Friendly by default: anyone should be able to work with Aave if the merits support it - Build and operate in public. Everything we do will be done openly, with the highest standard of accountability This is the direction we are committing to, a multi-year journey. The foundation is set. Now it's time to build. Aave will win.

English
0
0
2
127
Catalin
Catalin@dcatalin98·
@jxnlco the pipeline maps 1:1 tbh. "take a deep breath" is comfort building. "you are an expert at X" is pedestaling. "my career depends on this" is emotional dependency. prompt engineering is just the mystery method for LLMs
English
0
0
1
420
jason
jason@jxnlco·
we don't degrade models but if you've read anything about the PUA community its actually called negging.
English
15
0
178
18.9K
Catalin
Catalin@dcatalin98·
@Kekius_Sage because the world rewards confidence more than correctness. promotions, followers, funding, all of it goes to whoever sounds sure, not whoever's right. being confidently wrong has better expected outcomes than being uncertainly right
English
0
0
2
20
Kekius Maximus
Kekius Maximus@Kekius_Sage·
Why people are so confident when they're wrong?
English
371
15
330
18.8K
Catalin
Catalin@dcatalin98·
@aidan_mclau and you can't fake it. people who love the tech vs people who love the idea of working in tech. that gap shows up fast in the details nobody thinks to check
English
0
0
1
57
Aidan McLaughlin
Aidan McLaughlin@aidan_mclau·
it's really hard to compete with someone who earnestly loves technology
English
32
28
553
21.7K
Catalin
Catalin@dcatalin98·
@pmarca the neumann $350m check after wework was the pilot program
English
0
0
2
167