kevin

@k3vin-wang

153 Following

37 Followers

Colin pfp

We’re launching Coins on Paragraph next week - a new way for readers to support the posts they love & get rewarded, while helping great writing get the visibility it deserves. For writers: this means an entirely new way to earn from your best ideas. No paywalls, no gating, & no recurring subscriptions. Just instantly tradable, fairly launched ERC20s backed by your writing. DM for early access, rolling out today. See it in action: https://paragraph.com/@aja/the-death-of-the-yeoman-farmer

53 replies

94 recasts

475 reactions

kevin pfp

I believe tokenized stocks are a game-changer — why isn’t anyone talking about them?

2 replies

0 recast

1 reaction

kevin pfp

Is DeFi a common topic of discussion here?

1 reply

1 recast

0 reaction

sidhant pfp

🚀 just dropped the first-ever gemma 3n model running fully on ios — yes, on-device, no cloud. 📱 it’s slow (for now), but it benchmarks nearly on par with claude 3.7 sonnet for non-coding tasks. 🧠 gemma 3n is google’s new mobile-first frontier model, designed for edge devices. 🔧 i built a simple ios app to run it locally: 💬 features: on-device inference, real-time chat ui, streaming responses. ⚠️ it’s a bit slow, but it’s a start. 👀 try it out and see the future of on-device ai. https://github.com/sid9102/gemma3n-ios

3 replies

5 recasts

26 reactions

Daniel Barabander pfp

Daniel Barabander

I’m thrilled to announce our seed investment in Pluralis Research, which is building a decentralized AI training protocol. We are laser focused at investing in projects that use crypto ownership to solve coordination problems. Pluralis is a bullseye on that thesis, using ownership to solve the "resource problem" of training foundational AI models that are truly open. We couldn't be more excited to support Alex and the rest of the team. Full post here: https://blog.variant.fund/investing-in-pluralis-research

2 replies

36 recasts

97 reactions

kevin pfp

State of Crypto Report 2024: New data on swing states, stablecoins, AI, builder energy, and more https://a16zcrypto.com/posts/article/state-of-crypto-report-2024/

0 reply

0 recast

0 reaction

kevin pfp

After Uniswap released its Layer 2 chain, Andre Cronje pointed out that app chains are severely underestimating infrastructure and compliance costs (such as blockchain explorers, custodians, exchanges, oracles, bridges, toolkits, IDEs, on/off-chain channels, native issuance and integration, regulation, and compliance). He mentioned that in 2024 alone, these costs have already reached $14 million, with most of them being recurring expenses. Personally, I believe that from a consumer perspective, chain abstraction or interoperability will become increasingly important. Except for chain nerds, no one will care which blockchain this service is running on.

0 reply

0 recast

0 reaction

Jesse Pollak pfp

@jesse.base.eth

spent the afternoon at @nounstown.eth the /nouns ecosystem is such a prime example of builders doing good and having fun onchain

15 replies

58 recasts

201 reactions

🌈 YON pfp

yesterday was incredible. what i love about the nouns community is the unmatched creative energy everyone brings. here’s the skateboard i painted live, and a shirt design people got screenprinted there. blessed to have participated in /nounstown — thank you for having me!

20 replies

14 recasts

75 reactions

kevin pfp

David Perrel 一直都是我很欣賞的 Writer AI 可以幫你整理的很有條理但是寫作的靈魂和個性是需要日積月累的需透過閱讀經典作品來提高文字的感受力尤其對這句話特別有感 Writing Examples is the opposite of Grammarly. It celebrates the wild, wacky, and the weird because it’s the bedrock of personality. Writing Examples 與 Grammarly 恰恰相反。它讚頌狂野、古怪和奇特的表達，因為這些正是個性的基石。 https://writingexamples.com/

0 reply

0 recast

0 reaction

kevin pfp

基於 Base 和 Zora 協議的Tech Stack來打造客戶端應用程式是明智之舉。 song.camp 的音樂很好聽，打算在zora mint 一張收藏 https://song.camp/collection/0x2fAf66638A9D52187EfFec0b88155131FbB5098a?tokenId=1&chain=base

0 reply

0 recast

0 reaction

kevin pfp

https://blog.tulsk.io/p/crypto-growth-why-retention-is-key

0 reply

0 recast

0 reaction

kevin pfp

Here’s what I’ve learned from analyzing hypersub.xyz: • Prioritize retention and user experience (UX) over protocol in the early stages. • Retention fosters loyalty, reduces customer acquisition costs, and ensures sustained revenue. • Slower growth post-peak indicates a need for stronger retention strategies. • Enhance UX, personalize engagement, and continually provide value for both creators and users.

0 reply

0 recast

1 reaction

Zach pfp

Yesterday was a day I'll remember for a very long time. Thanks to everyone who made the first FC Dev Day so special! https://paragraph.xyz/@zd/fc-dev-day-reflection

5 replies

21 recasts

59 reactions

kevin pfp

Starting a conversation with OpenAI's advanced voice mode has been quite helpful in clarifying some thoughts.

0 reply

0 recast

0 reaction

Jason Goldberg pfp

I may be the person who’s most excited for Channel Memberships Here’s why👇 I’ve long argued that the current iteration of channels feels more like topic tags than Communities. You choose a topic when casting and then your cast goes to people interested in that topic. Community is different. Community is a group of people organized around common goals. Layer on onchain and it’s: common goals + collective ownership! A big part of our vision is that every channel member is a member-owner with skin in the game. When you come to a channel community in a couple weeks you’ll be prompted to buy a membership via Fan Tokens — which gives you access to the channel, partial ownership of it, AND shared participation in all rewards earned by the channel. This is the way.

25 replies

19 recasts

66 reactions

kevin pfp

The overall error of multiple models is equal to the average error of a single model minus the diversity of the model’s predictions. In other words, if you have many different predictions, the overall performance may be better than relying on a single prediction. The theorem can be expressed as a mathematical equation: {Average Prediction} - {True Value})^2 = {Average Single Model Error} - {Prediction Variability} The left-hand side represents the error of the overall prediction. The first term on the right-hand side is the average error of a single model. The second term on the right-hand side represents the diversity (or variability) of the model’s predictions. The theorem does not imply that any collection of diverse models will be accurate. If all of the models share a common bias, their average will also contain that bias. The theorem does imply that any collection of diverse models (or people) will be more accurate than its average member, a phenomenon referred to as the wisdom of crowds.

0 reply

0 recast

4 reactions

kevin pfp

This theorem has implications for decision-making in various fields: • Ensemble learning in machine learning: Combining multiple models can improve accuracy. • Democratic theory: Provides a mathematical argument for the wisdom of crowds. Limitations: • The theorem assumes independence, but in the real world, this may not hold due to shared information or mutual influence. • It assumes that each individual is more likely to be correct than wrong, which may not always be true. • In practice, increasing the group size might introduce coordination problems or dilute individual responsibility. Assuming each individual has a probability p > 0.5 of being correct: • The probability that the majority is correct is higher than p • As the number of voters n approaches infinity, the probability of a correct majority decision approaches 1 • As n increases, the chances of a majority error decrease • As n increases, the binomial distribution of correct votes becomes more concentrated above 50%

0 reply

0 recast

0 reaction

kevin pfp

The Condorcet Jury Theorem is named after the 18th-century French philosopher and mathematician, Marquis de Condorcet. It is related to the broader ideas of collective intelligence and social information aggregation. This theorem has been influential in discussions about the epistemic reasons for democracy and the design of decision-making bodies. When a group of individuals independently makes a decision between two options, and each person is more likely to be right than wrong, majority voting is generally correct. As the size of the group increases, the likelihood of the majority being correct also increases. It applies in the following conditions: • There is an odd number of voters • Each person has a greater than 50% chance of being correct • People make decisions independently • There are only two possible answers (right/wrong) As the number of people in the group increases, the accuracy of the majority vote approaches 100%.

1 reply

0 recast

0 reaction

Jesse Pollak pfp

@jesse.base.eth

thank you @ethereumfilm for creating this incredible piece

0 reply

10 recasts

26 reactions