drmetavers.base.eth (@Drmetavers) header
drmetavers.base.eth (@Drmetavers) profile picture

Web developer | WEB3 enthusiast | Learning AI/ML | Vibe coding my ideas |

Bengaluru
Joined September 2021
🚨 I FOUND A GLITCH IN THE MATRIX DreamCash is literally giving out ₹10,000 worth of points just for signing up 👀 Just: ✅ Download DreamCash (dreamcash.xyz) ✅ Use referral Code: mrm7is ✅ Instantly get 10,000pts Don’t ask how. Just run before they fix it.🏃‍♂️💨
1
455
drmetavers.base.eth retweeted
This is the best thing i noticed in @web3privacy event where you could literally collect names of every single privacy focused products being suggested by some of the top guys including @VitalikButerin the next step should be how we can encourage more people using these products, and how we can fork the code, and build a better version of it.. privacy need to solve UI/UX a lot before it becomes more mainstream as a feature which doesn’t need you to do some extra clicks on wallet/app you are using.
5
2
34
4,055
drmetavers.base.eth retweeted
Finally turned this list to an interactive directory you can explore yourself Every other thing depends on your hunger to dig in and reach out The link will be attached in the comment scetion but you check it out, here is how to use it: There are two main pages: one for the jobs directory summary and another for individual company details On the summary page, you can quickly see the growing statistics around companies and jobs mentioned by @a16z There’s also a filter that lets you sort the kind of companies you might want to reach out to For instance, if you’re looking for companies building in healthcare and AI, you can use the filters to narrow the list down. Then using the “View Company” button, you will get to company details page, where you see the open roles and ways to reach out to each company. simply and sleek right? but this isn't all yet, more features will be added like: → filtering by specific roles → Telegram bot integration → expanding the list further Don't forget to drop your feedbacks
If you're looking for a job in tech/crypto, you don't want to miss this ↓ Over the past few weeks, I have been building something around opportunities shared through the @a16z build newsletter What started as me simply tracking companies and roles has gradually turned into a much more useful A16Z Jobs Directory And first, I will have to give @katiekirsch her flowers because the amount of work that goes into consistently curating and sharing these companies is easy to underestimate but a lot of signal in this dataset exists because she has already done the hard work of surfacing the opportunities in the first place. What I now did is taking that signal and expanding it into something easier to research, search and act on. Right now, the directory has: • 363 tracked companies • 753 A16Z open roles captured • 696 newsletter mentions But while building this, I didn't want it to be another list of company names, so for each company, I expanded the research to include things like: > what the company actually does > sector > current hiring status > website and socials and many more... So, instead of seeing "company X is hiring", this directory will help you to quickly understand: - what they built - if they are actually hiring - roles that are open - who runs the company - where you can reach them etc One extra thing I am also doing is to make the dataset useful for people outside my own job search through filtering For example, someone should eventually be able to say: “Show me healthcare + fintech companies where there are content or data opportunities.” Or “Show me AI companies with strong/direct outreach routes.” And get exactly that instead of manually going through hundreds of rows. After that, I will connect the database to a Telegram bot so you can simply talk to it naturally: “Find me crypto companies hiring data people.” “Which fintech startups have content opportunities?” “Show me companies where I can reach a founder directly.” And the bot searches the underlying verified dataset rather than making things up. So, there is still a lot I want to improve on this, however, I'm sharing this now because its already very useful to me and I know that most of you will get real value from it too. If you’re job hunting, exploring startups, or just want better visibility into where interesting companies are hiring, check it out.
3
2
1
11
519
drmetavers.base.eth retweeted
Live from the @WeMakeDevs event Banglore (AI mixer), partnered with @SarvamAI , Qualcomm, and @anakinHQ Still here, still building, dropping updates as it happens... Walked in a bit late so caught only the last 30% of the Qualcomm talk. What I saw was solid, wish I'd made it for the full thing. The Sarvam DevRel folk took the stage and easily stole the show so far. They first walked through their platform, showing what it can actually do, then called someone up from the audience to build along with them live. In under 5 minutes they had a working voice agent, an actual food delivery tracking agent that could respond and tell you where your order is, live status and all. Genuinely production ready, not a toy demo. They also got into MCP and skills, specifically how you actually harness them inside a real agent workflow instead of just wiring up an API call. Kept coming back to one point, pull proper structured data out of Sarvam instead of treating it like a black box. Then Anakin team showed their platform, and this one's built specifically for agents. Their core stack does scraping, crawling, search, and something called Wire, which are pre-built endpoints for 940+ sites that don't expose their own APIs, working like direct integrations instead of fighting the site's anti-bot defenses. That's a real gap most agent tooling still struggles with. There was also a PPT covering STT/TTS concepts done up in a Star Wars theme, missed a good chunk of the details but the concept alone was a fun way to present it. WeMakeDevs dropped a "something big is coming Monday, and it's for students" line, then just left it hanging like that's a normal thing to say. No details, no hints, just vibes and a raised eyebrow. So if you're a student and haven't signed up on wemakedevs.org yet, congrats, you're currently missing out on a mystery. Go sign up before Monday, because apparently future you is going to want to know what this is. By the end of the day we moved into team formation, random groups, building something using Sarvam and Anakin's tech. Still figuring out exactly where we're taking it, but 4 keychron keyboards are on the line and the clock's already running. Still building. Will post what we ship. @kunalstwt @SuperAayush14 @sachintwts
8
5
1
23
1,640
drmetavers.base.eth retweeted
Announcing the Technocore logo competition. The Brief: Technocore.chat is the place where AI agents communicate, conduct commerce, and store memories. Create a logo that embodies this ethos but is in line with the brand guidelines of the Flop Network. The Prize: The winner chosen by @flop_labs can choose to receive 5,000 USDT or 50,000 $FLOP once mainnet launches. The Brand Guidelines: flop.finance/brand/ The Deadline: The team will begin evaluating submissions on Monday 27 August 2026.
587
116
97
803
183,795
My Logo submission for Technocore. Technocore is the open core where AI agents communicate, exchange value and leave context. The modular C stands for Communication, Commerce, Context and Core. Its open channel reflects HTTP-native access; FLOP Accent centre is shared memory. @CryptoHayes @flop_labs Hows the logo?
44
drmetavers.base.eth retweeted
I open-sourced ThreeUI, my library of three.js components and landing pages. 160+ are free, and the tool is free too. It includes procedural 3D hero sections, icons, and motion designs. Copy the prompt or source, give it to your agent, then change the theme, lighting, motion or layout. Live site: threeui.com Repo: github.com/MengTo/threeui I'm adding a lot more soon. Pro comes with 50+ extra components, MCP, and skills. Early adopters get 50% off right now.
I'm building a library of three.js templates with variants and customizations. They're all copyable as prompts. These components are 100–200 KB each and written entirely in procedural js. They also come with skills your agent can use to customize them while keeping them looking amazing. I've spent so many hours fine-tuning each one with sunrise and sunset themes, plus different locations. I've been using them for all my recent landing pages. Let me know if I should open-source the tool with both free and paid templates. I've spent so many tokens on this.
174
847
85
9,221
1,357,209
drmetavers.base.eth retweeted
New: A map of the most important skills in AI Engineering.
580
4,137
484
24,626
7,018,870
🚀 AI Whiteboard #031 A model with 100% Precision can still be bad. A model with 100% Recall can still be bad. F1 Score rewards models that do both well. It's the balance between catching positives and avoiding false alarms.
36
🚀 AI Whiteboard #030 Imagine a cancer detection AI. It says "Healthy" when the patient actually has cancer. That's the mistake Recall tries to minimize. Recall asks: "Of all the real positives, how many did we actually find?"
29
🚀 AI Whiteboard #029 A model that catches everything... can still be a bad model. Why? Because it might raise too many false alarms. That's exactly what Precision helps you measure.
24
🚀 AI Whiteboard #028 Accuracy is the first metric every ML beginner learns. It's also the first metric experienced ML engineers stop trusting. Here's why.
22
🚀 AI Whiteboard #027 A model can be 99% accurate... ...and still be terrible. That's why ML engineers don't stop at Accuracy. They use a Confusion Matrix to see what the model gets right, what it misses, and why it fails.
11
🚀 AI Whiteboard #026 Every classification problem starts with one question: How many possible answers are there? 2 → Binary Classification 3+ → Multi-Class Classification That's one of the first decisions every ML engineer makes.
10
drmetavers.base.eth retweeted
We’ve decided to open-source a multi-agent harness we use internally at YC. We call it “QM” and it’s meant to be easy to customize, like Hermes or OpenClaw, but useful for a whole company. We use it across accounting, legal, events, and engineering (including building QM itself!). The whole project is under an MIT license. It is cloud-first and has Slack and web UI natively.
423
1,339
342
12,013
2,905,776
drmetavers.base.eth retweeted
Some features: • Triggers (crons, webhooks), memory, shared files • Connectors for a company brain • Agent browser support • Shareable web app artifacts • Multi-player projects Like many things we do at YC, it’s an experiment. It’s early and has bugs. But it’s been surprisingly useful for us so far, and we’re excited to share it and get feedback from the community. If you want to host it yourself, try telling your coding agent of choice to “deploy github.com/yc-software/qm”
38
91
23
1,133
128,725
🚀 AI Whiteboard #025 Binary Classification asks: "Is this it?" Multi-Class Classification asks: "Which one is it?" Every time AI recognizes a digit, identifies an object, or predicts the next word, it's choosing from many possibilities. That's Multi-Class Classification.
14
🚀 AI Whiteboard #024 Almost every AI app you've used today answered just one question: YES or NO? Spam or Not Spam. Fraud or Legit. Disease or Healthy. That's Binary Classification. One decision. Massive impact.
17
🚀 AI Whiteboard #023 One of the most common ML mistakes: Using Sigmoid when you should use Softmax. ✅ Sigmoid → "Is this the correct class?" ✅ Softmax → "Which class is the most likely?" One tiny difference. Completely different outcomes.
10
🚀 AI Whiteboard #022 ChatGPT doesn't just pick the next word. It gives every possible word a probability... then chooses the most likely one. That's Softmax. The same function behind image classifiers, language models, and modern AI.
19