Rohit (@cashew_codes) profile picture

Building interesting things.

Joined October 2016
Photo by Rohit (@cashew_codes)Photo by Rohit (@cashew_codes)Photo by Rohit (@cashew_codes)Photo by Rohit (@cashew_codes)Photo by Rohit (@cashew_codes)
Ready to scoop $COOKIEs out of the $COOKIE jar?🍪 🔔The $COOKIE IDO allowlist is now OPEN on @Polkastarter! Apply here for a chance to get early access to the Cookie ecosystem, AI data layer, and multidrop access: allowlist.cookie.community  Details in thread! 👇
716
5,765
249
3,651
115,790
🔴RED DOT !! Accumulation phase has ended: no more easy buying opportunities in orderly and slowly increasing markets. Bull market has started. If history is any guide, we will see ~10 months of face melting fomo: extreme price pumps combined with multiple -30% drops. Enjoy!🚀
1,094
5,810
866
26,442
2,640,740
Why are people filled at 101.5$ $SOL if there is higher price order available on book. @DriftProtocol ? IMO drift algo will always give you worst possible price. You should NEVER use market orders only limit.
17
5
1
44
10,191
Sick of missing out on Solana airdrops? Us too! So we built a handy Solana airdrop tracker which we plan to update weekly docs.google.com/spreadsheets…
49
83
9
444
57,964
I remember listening to @tushar_jain on a podcast about a year ago. He said that every crypto cycle had its own novel token distribution mechanism. ICO in 2017, IDO in 2019 and now we have IPO in 2023. Points is the new token distribution mechanism of this cycle whether we like it or not.
7
4
2
71
10,815
Rohit retweeted
1/ Happy Holidays! 🎅🎉📚 The Aptos Foundation is bringing you a month-long Ecosystem Fundamentals campaign, starting today on Galxe. Learn and earn with Aptos today 👇 galxe.com/aptos/campaign/GCd…
1,477
100,974
310
20,381
1,796,102
Ready for the #NewYearsChallenge? 🎉 Here's everything you need to know to become the top noise collector! Dive into the details below 👇 #NoiseHunters #DePIN #Silencio
138
459
53
551
53,942
Rohit retweeted
1. Last week we announced the launch of $JTO, with 10% of the supply being airdropped to our community members. Today we’d like to share the details of airdrop eligibility and allocation specs. (we also wrote an in-depth blog post linked in tweet #2.)
282
334
218
1,415
696,832
7,735
18,218
5,325
143,809
34,148,658
“The purpose of software engineering is to control complexity, not to create it.” - Pamela Zave
33
540
41
2,649
212,897
Wall Street is coming to #Bitcoin Bullish 🚀
422
1,761
92
5,965
598,854
Rohit retweeted
this time last year, @heliuslabs didn't exist in a short time frame, with significantly less resources than our competitors, in the bottom of a brutal bear market — we've been fortunate to support many of Solana's finest teams build amazing experiences that are only possible on Solana from APIs & RPCs to webhooks & hackathons, we've been able to do all of this due to a single reason: relentless focus on Solana developers the goal of Helius isn't to just give you the best RPCs in the ecosystem, it's to help you succeed on Solana — whether that's 24/7 support, random AWS troubleshooting tips, or even just marketing help we simply have a deep love for the game and we're just getting started (seriously), we'll be here building with you until the end here's to everyone building the future of crypto on Solana and thank you to everyone for the support so far 🍻
36
48
9
363
50,202
1/ @Solana vs the field: Modularity proponents doubt Solana's scalability, but Solana believes its integrated design offers advantageous benefits. The question is: Does Solana meet the cost requirements of its customers? And is it gaining traction? @kelxyz_ dives in. 🧵
32
118
14
433
160,402
GM twitter. Time management is underrated among us adults. How do we teach kids about it now so it may be better when they will be adults?
1
2
82
✨ I spent the last 48 hours making GPT-4 read the entire @solana validator codebase and write documentation, so @toly doesn't have to. Introducing TolyGPT.com — an AI-powered chatbot trained on nothing but code that can answer deep technical questions. How it works 👇 But first... A huge shoutout to @langchain, @chillzaza_, and @_seanyneutron. Their hard work made prototyping this thing a breeze. Without further ado... Devs like to write code, not documentation. Tribal knowledge is lost when devs move on to other projects, leaving future devs to sort through mountains of code and figure out not just how it works, but why it works that way. This is all about to change. GPT-4's ability to write code is stunning. It seems to understand something fundamental about writing software that previous models just didn't. This comprehension of the principles that drive the design behind a complex system carries over into its ability to document existing codebases in a truly impressive way. With the enlarged context window(s), it's now feasible to feed GPT-4 entire files of code and ask it to write documentation about how the code works. Taking this as a starting point, the process looks something like this: 1. Download repo. 2. Depth-first traversal of repo contents, ignoring things like package-lock and binary files. 3. For each file, feed to GPT-4 and ask it to write documentation in markdown. 4. Save the output in a separate location as [outputRoot]/[inputFilepath][inputFilename].md 5. For each folder, we ask GPT-4 to write a summary of the folder, taking the newly generated documentation for all files in the folder and the summaries from each of its subfolders as context. Write this to the filesystem as markdown. Now we have a filesystem that matches the structure of the input repo, but all files in the tree are markdown documentation of the corresponding code file. From here, we: 1. Load markdown documents into LangChain. 2. Embed all documents via @OpenAI embeddings. 3. Store embeddings in @pinecone. When a user sends a query: 1. Embed query. 2. Find k-nearest markdown files. 3. Feed to GPT-4 with a prompt asking to answer the query based on k-nearest markdown documents provided. The craziest part of all this? GPT-4 actually wrote ~30% of the code. The results are pretty good for 2 days of work. There is certainly room for improvement. Some items that are top of mind: 1. TolyGPT will occasionally hallucinate answers. It is especially bad with links to external sources, like GitHub. The base model seems to know a bit about Solana already, and sometimes this creeps in. Fine-tuning the prompt can solve some of this. 2. Context selection is difficult in a codebase this large. For example, sometimes it will pull in details about the Solana SDK when asked about transaction processing. The SDK files can seem relevant depending on the phrasing of the question. It may be worth breaking the documentation into subsystems to limit this. 3. Not all files fit into the 32k token window. As of now, there are 23 (out of ~1,100) files that cannot be documented in their entirety. Some of these files are very important to how Solana works. Final thoughts: 1. GPT-4 is super powerful, and we're going to see a ton of tools that supercharge the entire software development lifecycle. This is not 12 months away. For the people that can afford it, these tools are here now. And they're only getting better. Act accordingly. 2. The price of inference has to come down for this to go mainstream. I spent about $300 prototyping this project, and the final crawl cost about the same. The high cost of GPT-4 will push developers to other, cheaper alternatives with similar performance. This is coming very soon. If you have a large software project and you're interested in something like this for your codebase, fill out this form tally.so/r/nr502L, and we'll be in touch this week. Or just DM me :)
84
225
57
1,309
374,742
Helium is coming to Solana ❤️‍🔥
1/ An update on the @Helium Network migration to Solana: the Readiness Working Group has voted to shift the migration date to April 18 to provide more time for community readiness, necessary governance votes, more testing, and operational improvements. medium.com/helium-foundation…
3
23
108
6,818
How to fall asleep quickly. The Military Method (works for 96% of people):
489
6,036
465
30,029
Metamask will now track your IP address for every transaction you send Here is their latest privacy policy: consensys.net/privacy-policy… There are ways to avoid getting tracked, follow on... 1/7
81
384
51
913
1/ Curve just released the Curve stablecoin docs. The biggest innovation is Lending-Liquidating AMM algorithm (LLAMMA). So, how does it work? *I'm learning as I'm writing this thread*
63
381
56
1,648
Having a safe CEX: proof of solvency and beyond vitalik.ca/general/2022/11/1… Big thanks to @balajis and staff from @coinbase @binance @krakenfx for discussion!
1,002
1,680
346
7,396