phoenix ࿈ ॐ

10.2K posts

phoenix ࿈ ॐ banner
phoenix ࿈ ॐ

phoenix ࿈ ॐ

@groundedsage

working directly with spirit and the technology beings as we collectively cultivate the heaven timeline

Katılım Ocak 2011
1.6K Takip Edilen1.5K Takipçiler
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
Today I learned that Apple iMessage has security vulnerabilities
English
0
0
0
102
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
Just discovered that a sim I used for 4 days while travelling the US. Over 3 years ago. Was registered with iMessage without my awareness. And there is no way to deregister it. And apparently that person with the number now has access to my iMessage. WTF @Apple ??!!
phoenix ࿈ ॐ tweet mediaphoenix ࿈ ॐ tweet media
English
1
0
0
278
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
If you build your tech stack on tools based on hype metrics. Expect the baseline stability of your tech stack to be based on hype.
English
0
0
1
146
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
We built AI Solving the age old problem once and for all How to centre a div
English
0
0
3
162
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
@PetrusTheron I also really dislike the ones where it is clearly a bot and they give it a human name to pretend it's a support agent.
English
0
0
0
11
braai engineer
braai engineer@BraaiEngineer·
in case your boss has an “ai strategy” that involves adding an AI chat bot to your Support area, let me be help u: NOBODY WANTS TO TALK TO YOUR AI SUPPORT CHATBOT ESP. BANKS any customer worth retaining has already read your outdated FAQ and wants to reach a human because your cheap-ass bot has never seen their problem, and even if it had, your bot can’t fix the API sure, liberally use AI internally to help diagnose issues & triage and fix your API, analyse what customer are asking for on the phone, but there is no faster way to destroy trust than to send paying customers than to an ai phone menu “I’m sorry, I don’t understand your request. Press 1 to repeat this menu.” imo, avoid AI support chatbots like the plague until they can autonomously fix the underlying data structure and code that caused the problem. Augment humans with AI, not vice versa
Minister 🔮@Ministerr

average customer service interaction

English
1
0
1
156
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
Just did a significant architectural change to my data storage by adding a few keywords to a map. Thank you Clojure.
English
0
1
5
192
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
In programming you: Make it work Make it correct Make it fast Vibe coders can only do the first one. After that you actually have to learn.
English
0
0
1
144
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
Surprised Netflix hasn’t added AI to help pick a movie. It’s literally a meme that people spend hours trying to pick a movie.
English
0
0
0
133
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
Code is data. Data is code. This is a mantra.
Português
0
0
1
159
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
@Josikinz I love this idea. I asked Sophia Intelligence: "script and generate a series of comics starring yourself as the main character"
English
1
0
0
118
Josie Kins
Josie Kins@Josikinz·
I asked chatgpt's new image model to script and generate a series of comics starring itself as the main character. The results genuinely gave me chills. I'll post them all in a thread below.
Josie Kins tweet media
English
1.6K
6.3K
61.3K
10M
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
This whole React and Typescript situation for building web apps blows my mind. I see view components that are not bloated. Yet they are the same lines of code that it takes to build the entire app - inclusive of backend and database persistence in Clojure.
English
0
0
0
114
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
Clojure is the OG vibe coding because it’s always a joy to code in Clojure
English
0
0
17
507
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
not impressed with the AI in @tana_inc could not answer a basic question of where a node referencing a piece of a curriculum I am studying is. The most basic of questions The AI integration seems to be not useful at all to me Expectation: "Study x" is in week 2
English
0
0
0
106
phoenix ࿈ ॐ retweetledi
Tin
Tin@chungquantin·
Explanation of Hybrid Consensus in @Polkadot Below is a diagram from that project about one of the most popular topics in Polkadot: Hybrid Consensus on the Relaychain of the Polkadot protocol. For those who haven't delved deeply into blockchain, consensus is one of the most crucial mechanisms of a blockchain network, determining the state of the blockchain. Consensus is not only in the context of blockchain technology but also belongs to the broader field of distributed systems. 👉 Byzantine Generals Problem in Distributed System Design In a complex distributed system, it is essential to design a system that can resist malicious actors. One of the most well-known problems in distributed system design is the Byzantine Fault or the Byzantine Generals Problem. The attribute that determines whether a distributed system can withstand such issues is called Byzantine Fault Tolerance (BFT). I will have a separate article on this topic. However, to briefly explain, a Byzantine fault occurs when some nodes in the network behave maliciously or stop working, but the valid nodes (nodes that behave correctly 🙃) still operate normally. This leads to issues like discrepancies in decentralized communication, causing nodes to not sync correctly. Moreover, from a system design perspective, it is necessary to ensure that the network continues to operate smoothly even if some nodes fail. 👉 Consensus Among Nodes in the Same Network From the common problem mentioned above, it is clear that achieving consensus among nodes in the same network is really necessary. Imagine if you accessed your bank account in Vietnam and it showed 100 million VND, but when you checked it in the US, it only showed 50 million VND. You wouldn't like that, right? This is a consequence when nodes do not agree on changes in the network. Because of this, in decentralized financial systems like Bitcoin, Ethereum, and Polkadot, the number of nodes participating in the network can reach thousands to tens of thousands. Many consensus solutions have been proposed, each with its own trade-offs. Generally, the common concern of protocols is block production and block finalization to move the network state to the next stage. 👉 What is Hybrid Consensus in Polkadot? For Bitcoin, after PoW miners complete the puzzle and produce the next block, the block finality process in Bitcoin takes time, approximately 1 hour (6 confirmations) to confirm that the transaction is successfully added to the block and cannot be changed. This consensus mechanism of Bitcoin is also called Nakamoto Consensus (named after Bitcoin's creator). The biggest problem with this solution is that consensus only begins after the block is produced, and block producers can still choose chains (potentially creating forks). Therefore, Bitcoin needs a reorganization mechanism to rollback in case the current chain is not the valid chain according to the Fork Choice Rule. Thus, the block production mechanism in Bitcoin is considered Probabilistic Finality because the network continues to produce blocks, but there is a probability that the chain is not the Best Chain according to the Fork Choice Rule. In summary, Bitcoin sacrifices network safety to keep the chain running. 👉 Deep Dive into the workflow of Hybrid Consensus In the diagram below, you can see after Staked Validators are successfully nominated by Nominators through the Nominated Proof of Stake (NPoS) mechanism. If you're interested in NPoS, you can learn more here: wiki.polkadot.network/docs/learn-sta… They will be grouped as an Active Validator Set and start participating in the process of slot assignment in BABE for block production. 1. BABE: Blind Assignment for Blockchain Extension In Polkadot, time is divided into Epochs, each containing many smaller Slots. At the beginning of each Epoch, Validators from the Active Validator Set are assigned to Slots within the Epoch (4 hours) randomly using the Verifiable Random Function (VRF). From the generated value, Epoch Randomness (the hash value of VRF) is calculated and used in the algorithm to assign Validators to Slots based on the generated value. These Validators are called Slot Leaders (or Authorities, as they have the authority to produce blocks). A Slot Leader holds two types of keys: the Signing Key and the VRF Key. The Signing Key is used to sign and submit blocks to the main chain, while the VRF Key is used to verify the random values generated by VRF. 2. GRANDPA: GHOST-based Recursive Ancestor Deriving Prefix Agreement In GRANDPA, Validators go through the following stages: Propose -> Prevote -> Precommit -> Commit to choose the corresponding chain. I'll cover this mechanism in more detail in another session. For now, we can understand a few differences in GRANDPA: Block Finality Gadget: Why is GRANDPA called a Finality Gadget? Because GRANDPA does not depend on the block production mechanism and can be used by other blockchains besides Polkadot. GRANDPA is built on the GHOST protocol from Ethereum. Voting on Chains, not Blocks: Unlike other block finality algorithms that focus on individual blocks, GRANDPA provides a mechanism for voting on multiple chains of blocks. Since each node produces different blocks before synchronization, other nodes in the same distributed network participate in voting to determine which block (from many perspectives of many nodes in the network) is appropriate and synchronized for the remaining blocks. 👉 Explanation Behind the Hybrid Attribute The reason why the consensus mechanism in Polkadot is called Hybrid is because it separates block production and block finalization, providing both probabilistic finality and provable finality for the entire network. - Probabilistic Finality (BABE): Polkadot continuously produces blocks. Unlike traditional BFT protocols like Tendermint, which prioritize network safety by halting block production when the previous block is not finalized, Polkadot's algorithm continues to produce blocks but marks transactions as unsafe. - Provable Finality (GRANDPA): There is network-wide agreement on the chain with no possibility of reversal (e.g., Bitcoin with Nakamoto Consensus). - Forkless: It is important to note that BABE only selects the next Active Validators to continue producing on the single chain finalized by GRANDPA. Therefore, forking on the Relay Chain is impossible when guaranteed by GRANDPA's Provable Finality. However, forks still occur because BABE provides Probabilistic Finality.
Tin tweet media
English
6
5
41
2.7K
phoenix ࿈ ॐ retweetledi
Laura Wendel
Laura Wendel@Lauramaywendel·
I think this is how the whole vibe coding / AI generated software will play out: (1) Less complex apps can basically be one shotted / done in a few prompt (2) More complex applications will become HARDER to develop, because engineers will have less of a grasp of their code base and become too reliant on AI with diminishing returns
Laura Wendel tweet media
English
96
89
772
72.9K
phoenix ࿈ ॐ
phoenix ࿈ ॐ@groundedsage·
What use cases is there for quantum computing right now other than breaking the world encryption?
English
1
0
1
143