How to Stay Safe in a DeFi Era

Bitcoinist发布于2022-02-16更新于2022-02-16

文章摘要

Decentralized finance (DeFi) came into existence a few years ago but only recently rose to fame.

Decentralized finance (DeFi) came into existence a few years ago but only recently rose to fame. Like cryptocurrencies, this emerging financial technology runs on distributed ledgers and replaces intermediaries with smart contracts, making the entire asset management process automatic. With a $127.06B total market cap and $78.92B worth of crypto assets locked in P2P protocols, DeFi rightly holds the trophy of the fastest growing sector in the crypto space.

However, decentralized finance is not immune to common problems. The sector suffers from the same old disease that has been plaguing the crypto industry for a decade – scams.

The amount of money lost in DeFi is overwhelming. According to London-based firm Elliptic, users suffered losses totalling $12B, out of which scams were responsible for $10.5B. The figures are horrifying yet don’t condemn the sector. DeFi has a lot of potential – projects like Aave do give hope for the future. To avoid scams and find truly robust DeFi solutions, it’s vital to do your own research.

Product Reviews

Reviews help us learn about any underwater stones that the service or product have from former or current users. The feedback has a great value – it saves time and energy and, in some cases, money.

However, things don’t look simple when it comes to the crypto industry. After all, countries don’t have proper regulation mechanisms, and information can’t be easily verified. Yet, emerging review sources like TrustPilot started to pop up recently, e.g. RugDoc. The site has a big community and encompasses reviews on 2,600 crypto projects – yield farms, exchanges, OHM forks, play-to-earn games, and NFTs – spanning across 24 networks, including Solana, Binance, Polygon, Avalanche, and Cardano. RugDoc also monitors new projects with Launch Calendar.

At the moment, the crypto review niche is in its early development, but a few more projects are expected to emerge soon.

“Rugs” Under Rugs

Most scams are plain simple rug pulls. In the first phase, fraudsters add liquidity (usually in the chain native currency) to their tokens and attract investors. Then, they remove liquidity and run off with funds, pushing the token’s value to near zero and leaving holders with nothing.

There are other ways to fool the crowd, and Squid Game Token attributed to the show of the same name is a good example. Fraudsters simply disabled the ability to withdraw tokens for investors, leaving them with an unsellable asset. Thus, it’s vital to assess the projects from a technical standpoint.

Some companies have already implemented a strict audit and a special rating system. For example, RugDoc ranks projects from least risk to highest risk through a technical review of the project’s smart contracts- a controlling program deployed to the blockchain. The company also has a KYC Program that uses the process known as “doxxing” to verify the team members’ identities to hold them accountable in case of malicious events.

DeFi Monitor Tools

There are many DeFi tools that monitor crypto projects. Their sole purpose is to identify blind spots not visible to the general public and allow users to invest safely and efficiently in decentralized protocols. Some offer customized audits (e.g., DeFi Audits). In contrast, others have open-source tools (e.g., RugDoc), such as honeypot checker to verify if smart contract vetoes token sales, emergency withdrawal to let users remove funds from the project, and LP Breaker to separate LP tokens.

DeFi Knowledge

DeFi is a rapidly evolving industry with new projects popping up every single day. However, it’s essential to dive deep into details. That way, users can avoid scams and invest with certainty. That’s why many crypto projects have launched their own wiki hubs. In these hubs, users can find all the necessary info about decentralized finance and learn more about the product functionality. Projects like PolkaDot or Ethereum all have wiki hubs.

Furthermore, it gives users confidence that the project is up and running. However, not all have a portal of that kind, which is a reason to become suspicious. Wiki hub, just like White Paper, is essential for community development and trust.

In some cases, the project simply doesn’t have enough resources to create an entire separate wiki hub. In this case, users can turn to publicly available wiki libraries, like RugDoc Wiki, which covers tutorials and information on all Ethereum virtual machine compatible chains. Public knowledge sources are created for the community by the community. These places unite both crypto pros and investors where they share up-to-date information.

Wrapping Up

DeFi is one of the fastest-growing sectors in the finance industry. It offers great potential in transforming the financial industry and equally as many threats due to its decentralized nature. Thus, it’s vital investors adopt a security routine to analyze risk and access sources with additional information on the product or educational articles. That way, any user will be empowered to stay safe in the era of decentralized finance.

你可能也喜欢

Deepseek 能否为中国节省 1 万亿美元?

《DeepSeek 能否为中国节省 1 万亿美元?》一文探讨了DeepSeek如何通过技术创新,可能大幅降低中国AI基础设施的建设成本。 核心在于,AI硬件的成本正从计算芯片向昂贵的内存(如HBM)倾斜。DeepSeek通过三方面技术突破,系统性降低了对这类稀缺硬件的依赖: 1. **压缩“记性”**:采用多头潜在注意力(MLA)等技术,将长上下文对话所需的缓存(KV Cache)压缩至原来的极小部分,极大节省显存。 2. **按需唤醒“身体”**:利用混合专家(MoE)模型架构,每次只激活部分参数,并结合智能的存储调度,将多数参数移出昂贵显存。 3. **复用计算结果**:将算过的上下文作为缓存复用,用低成本的内存读取替代高成本的重复计算。 这些技术协同,使DeepSeek V4等模型在长上下文场景下,用同等硬件可能产出数倍的有效Token,等效于大幅提升硬件效率、摊薄单位成本。 文章据此进行推演:到2030年,全球AI硬件投资预计达数万亿美元。若DeepSeek的技术路线能使中国未来AI基建的硬件需求等效减少75%,在日均Token消耗达到数千万亿级的规模时,节省的投资额可能接近1万亿美元(约7万亿元人民币),相当于少建数万座智算中心。 更重要的是,此举将产业价值从受制于人的尖端算力芯片,部分转移至中国已有所突破的存储芯片及系统工程领域,提升了供应链安全性。DeepSeek的意义不在于“消灭硬件”,而在于通过极致优化,降低行业对最昂贵、最稀缺硬件的边际依赖,让AI能力更普惠,重塑未来AI基建的账本。 (注:文中关于万亿节省的具体数字属于基于行业趋势的推演观点。)

marsbit38分钟前

Deepseek 能否为中国节省 1 万亿美元?

marsbit38分钟前

推翻主流治幻觉思路:元认知,才是大模型破幻觉的全新解法

谷歌研究院与特拉维夫大学联合发表论文,提出对抗大模型“幻觉”问题的新思路:与其追求让AI全知全能或过度拒答,不如培养其“元认知”能力,即让模型能够感知并诚实表达自身对每个答案的确信程度。 论文指出,当前主流方法存在局限:一味增加知识覆盖无法穷尽所有事实;而通过大幅拒答来降低错误率则会征收沉重的“实用性税”,牺牲大量本可正确回答的问题。核心原因在于模型缺乏“判别力”,难以精准区分具体问题的对错,导致校准良好但实用性低下。 论文重新定义了“幻觉”:问题不在于AI输出错误信息,而在于其“没有资格确定却以确定的语气给出错误信息”。因此,解决路径应是实现“忠实不确定性”——让AI语言表达的确信度与其内部状态的真实确信度对齐。这比消灭所有错误更可行,是一个依赖内部信号的闭环问题。 在AI代理(Agent)时代,元认知更为关键。没有它,Agent在调用外部工具(如搜索)时将陷入“盲飞”,无法智能决策何时需要搜索、如何评估信息可信度。 实现元认知面临几大挑战:“自举悖论”涉及用静态数据训练动态能力的困难;“对齐破坏信号”指RLHF等训练可能磨灭模型原有的内部不确定性信号;“因果性评估”则需区分真正的元认知与对其的表演。 论文建议,评估反幻觉方法应超越单一准确率指标,转而分析完整的“实用性-错误率权衡曲线”,并关注其在其他任务上的“附带损伤”。最终目标是让AI学会诚实地沟通其认知状态,从而在保留实用性的同时,将错误信息的危害降至最低,建立可靠信任。

marsbit43分钟前

推翻主流治幻觉思路:元认知,才是大模型破幻觉的全新解法

marsbit43分钟前

交易

现货
合约
活动图片