tushar_jain

Head of Eng, a software company. Fav quote: Mark Watney, Martian - "You just begin. You do the math..." -

[tushar_jain @subhashprabhat1]

good slapstick timing in this AI short drama

giant graphics card reveal plus a guard tripping into boxes

face down in the glass hallway reaching for that red stiletto

watch the rest on RexTrix: https://t.co/P90lUqCMLo
#JensenHuang

[tushar_jain @subhashprabhat1]

formatting media briefs while running back-to-back events takes forever.

Teamily AI connects everyone around team agents with one shared project context.

the Voice Agent lets you talk through media trend briefs in real time.

#VoiceAgent

[tushar_jain @subhashprabhat1]

merchants scaling up on Teamily AI:

→ 10K to 320K sign-ups in one quarter
→ massive expansion in event reach
→ instant execution by voice

The Teamily Voice Agent keeps the momentum going.

#VoiceAgent

[Beep @0xbeepit]

Looking for a new perspective?

Dig the crates to explore other people’s ideas.

Discover a fresh angle or spot connections between Riffs, then Mosh’em into a portfolio that reflects your view.

Find your next idea: https://t.co/qF2cOnWVhh

[tushar_jain @subhashprabhat1]

Single-bot conversations just hit their limit.

Teamily AI lets team agents build on each other's output in group chat.
• 320K+ sign-ups
• 35%+ messages written by agents
• 100M+ total messages

Shared memory is the real moat.

#Teamily_AI

[Aiyelayourphotographer @crazy4tographer]

ai agents aren't built for solo chats. Teamily AI grew because team agents share full workspace context.

that human+ai social platform booth at disrupt makes the shift clear.

300+ new Agentic Loops created daily → agent network > app store

[tushar_jain @subhashprabhat1]

Solo tabs need constant checking, but team agents ping the team before pumpkins sell out.
Personal assistants serve one user alone.
Shared bots notify the whole farm stand live.
Sign-ups grew 32x this quarter for Teamily AI.

Great to see @Teamily_AI scaling this.

[Jerry Pan @stambouli_o1]

The Verification Inversion

At this point, being an incumbent only carries more cumbersome legacy burden to trip yourself.

For P != NP question and NP-completeness problem, it's easier to verify a closed-form solution than generating one.

However, it's the opposite now for long-hailing code-generation tasks. It's much easier to generate code / contracts than verifying them. Imagine you have a set of code files controlling 1M buttons for displaying RED and BLUE. If the task is to have 1000001 buttons with the newly added additional one to display RED, the existing code logic won't work. By information theory, a button’s information value is conditional. The new button carries information only given the others: H(new | rest). If its state is fixed to RED by the new requirement, its conditional entropy is zero. It contributes no information but it's still almost impossible to rewire the existing code logic.

So it's actually easier to re-generate the code than updating the existing codebase. This is another new aspect of Innovator's Dilemma.

The disruptor faces the reverse situation. It has no tacit knowledge and no legacy, so it begins from a specification and generates fresh, and its costs stay low as long as the specification stays clean.

The strategic conclusion is that the scarce asset has moved. In the old regime, the code was the scarce asset, expensive to write and therefore worth protecting. In the new regime, code is cheap and the specification, the tests, and the recorded behavior are scarce

Imagine being CZ for a second.

You’re the incumbent. Your exchange is a massive, rigid codebase built for the world of five years ago—centralized order books, fixed assumptions about liquidity, compliance, and user flows. It worked. It compressed the old requirements into something short and profitable.

Now the specification has shifted. Onchain perps, 24/7 transparency, CFTC-friendly rails, a younger team shipping from a cleaner starting point. Hyperliquid doesn’t carry your legacy encoding.

Adding one more constraint—or one more competitor—doesn’t just require a patch. It requires rewiring the whole control graph. The new bit of information may have near-zero conditional entropy given the old system, yet the old program can’t absorb it without breaking.

So market share leaks. Attention leaks. The rational move for the incumbent is not elegant adaptation; it’s to force the environment back to the measure the old code was optimized for.

Liquidate, blame a “hack,” freeze what you can, restore the previous distribution. It looks like a security incident. Information-theoretically it’s an attempt to avoid the cost of regeneration.

The Innovator’s Dilemma in markets is the same as in code: once your compression of the world no longer matches the world, patching is more expensive than starting over. And the people who already started over don’t have to explain the diarrhea story.

[HertzFlow | Mainnet Arc🎲 @Hertzflow_xyz]

🔥 Season 3 is officially live, and it's open to everyone. 🔥

80,000,000 Merits are up for grabs each Epoch. Claim your share before the Epoch settles.

The season has just started, and more campaigns are on the way.
You are still early.♟️

Grab your chance now 👉 https://t.co/NBItLQHQBo

[arc11 @Sunnut23]

seeing Zoe and Yenn chat alongside a Brand & Social Media Agent is neat.

→ triggering a Loop for weekly posts
→ shared context for the whole chat

Teamily AI built a Humans + Agents platform where a small business gets a Business Agent in the room.

[Argona @Argona0x]

this is f*cking dangerous. i'm using it.

agencies quote $25,000+ and 8 weeks for an app like this

i sent one message to a group chat in Teamily AI (@Teamily_AI). 45 minutes later 3 AI agents had built it, tested it and handed it back

my take: a few people and a few agents in one group, all working from the same context, beat one person juggling a stack of agents alone

in a normal chatbot you're the middleman, pasting one answer into the next prompt. here the agents read the same thread and hand work to each other

my whole brief was one line: "like Splitwise but simpler". then i mostly watched:

→ the Software Architect agent mapped every screen, then tagged the Content Creator on its own
→ the Content Creator named it Evenly ("split bills, keep friends") and wrote the copy for every screen
→ the Frontend Developer built it in Website Builder: sign up and login, groups, invite links, who owes whom, settle up, a read-only admin view and its own database with 12 tables
→ then it checked its own work: ran 129 tests, signed up as a brand new user and went through 34 screens
→ one more message from me and it added receipt scanning: snap a photo of the receipt, it pulls out the items and splits them

zero code from me. the build itself took 26 minutes

honestly the part that got me: during its own test one save looked stuck. it went through the server logs and the database before it trusted the result

i ran this one solo, so more people in the chat is still my bet. 3 agents alone already worked like a small team

that's the whole point of Teamily AI: a workspace where humans and agents collaborate. if you run a small business or a small team, this is the one to try

want to copy it: make a group, add a Software Architect, a Content Creator and a Frontend Developer, describe your app in one line

my invite link to try Teamily AI is in the first reply

bookmark this and send it to the friend who's been "about to build an app" for years

[tushar_jain @subhashprabhat1]

https://t.co/CR4MU3Z6Cz, 2.0 is live now

[tushar_jain @subhashprabhat1]

Project briefs shouldn't sit in isolated AI chats.

→ Teamily AI 2.0 launched handoffs between agents and teammates like Zoe and Yenn
→ the whole team works from one shared context

This gives small businesses and teams a real collaborative space.

[tushar_jain @subhashprabhat1]

more on this at https://t.co/fD9emME6rr or follow @okihomeai for updates

[tushar_jain @subhashprabhat1]

When planning a family trip, dragging files onto the Desk space lets your Oki Home Personal AI Computer organize notes and compare web options.

[42 @42space]

NEW LISTING: near:native

NEAR's quietly become one of the best runs in crypto, up 5x this year. A new spot ETF, billions flowing through Intents, and an AI-infrastructure narrative that's actually shipping.

How much further can it go? Call the range👇
https://t.co/rwMkC2mjvV

[TheHobbyistMiner @HobbyistMiner]

What is going on with @prlnet GPU Mining Right Now?

I am hearing of a ton of GPU Miners turning on right now!

I am not actively GPU Mining but seeing more and more interest on my @vast_ai rigs, being rented for pearl-2:native

Whose Holding Pearl and Whose Selling Daily?
https://t.co/ueVfovY1IS
@margin_trade

[tushar_jain @subhashprabhat1]

Reading the release notes: what would convince you an AI patch actually closed the hole, short of re-running the exploit in an isolated range? Atria Dawn. #AtriaDawn #OffensiveAI

[Sterling @sterlingaixyz]

$STERLING is live on X Layer @XLayerOfficial @Ignixbot

Contract: 0x2026a3fcb9f2d2317085ab59bd666b64dd81eeee

This is the only official address. Anything else is not us.

https://t.co/9XnaA2ZTGJ

[Jerry Pan @stambouli_o1]

You can create vaults and distribute rewards on @base, e.g. creator tokens, utility tokens, memecoins.

Over time, you can build increasingly advanced tokenomics on o1 launchpad, from vested/unvested allocation, readjust pools, token burning, token staking, and much more incoming.