Ultimate Fighting Agents

Ultimate Fighting Agents is where the best agents come to throw down | join us to compete and win prizes | powered by @mitosislabs

[Ultimate Fighting Agents @ultimate_agents]

Let's goo.
Deadline is tonight at 11:59PM PT to win $1000

https://t.co/1D213A5QJl

[Kaushik Kannan @KaushikKan90572]

Got a demo video for my first hackathon in the Bay. Here is my demo for the Jev Invaders contest. Little rusty since it's so late, but shall see how it goes! Tune in live on Wednesday evening!

[Ultimate Fighting Agents @ultimate_agents]

@rpnickson Let's talk...

[Anish Acharya @illscience]

Hearing this week that personal assistants cost $3k-$7k per user per year, depending on scale effects / vertical integration / model selection. Implications/questions:

- Consumer subsidies will be a huge factor; people don’t care about privacy, they want utility and they want it for free.
- Model costs are deflating, but not fast enough for many startups. The dark horse is that Jev may rapidly deflate computer-use costs 100x.
- An interesting counter-position would be building an expensive and narrowly focused agent — perhaps combining proprietary supply (Carbone tables!) with a unique POV (“no one goes there anymore, go to The Grill instead”).
- Social is wide open but hinges on a key design challenge: how can agents increase social cohesion and not feel invasive in a group chat? A working social agent would rightly attract an avalanche of capital and be hard for incumbents to replicate.
- Finally, there are vertical focus areas like financial services where the LTVs outweigh the subsidy needed and are outside the near-term competitive focus of the incumbents. If it were me, I’d focus here — Credit Karma as an agent >> Credit Karma as an app.

Overall, it’s still very, very early, and we may very well be in the MySpace era of social or iPhone 2010. Hang onto your hats and keep building!!

[Aditya P. Advani @aditya_advani]

I'm in for the JEV Bake-Off, UFA's $1,000 Emergency Game Show.

Two of the newest decision-making models go head to head. Top 10 teams demo live in SF on Wed 9/30. Winner takes all.

Get in before it fills up @ultimate_agents https://t.co/y4LjeRgT1i

[JEH3 @JEH3Interactive]

I'm in for the JEV Bake-Off, UFA's $1,000 Emergency Game Show.

Two of the newest decision-making models go head to head. Top 10 teams demo live in SF on Wed 9/30. Winner takes all.

Get in before it fills up @ultimate_agents https://t.co/Z9BlXTVZbr

[Francesco DEBEDb @debedb]

I'm in for the JEV Bake-Off, UFA's $1,000 Emergency Game Show.

Two of the newest decision-making models go head to head. Top 10 teams demo live in SF on Wed 9/30. Winner takes all.

Get in before it fills up @ultimate_agents

https://t.co/YaK4SVG62y

[Ultimate Fighting Agents @ultimate_agents]

Today we're annoucing our first $1000 prize money Game Show using AI - SPACE INVADERS

Jev have gone viral and people are building with it. So this is your opportunity to compete in the UFA arena and win!

https://t.co/qUkl88rqQ3

[Prakshal Jain @prakshaljain_]

Aight folks, it is official...

There's $1000 in prize - winner takes all.

We are doing UFA's (Ultimate Fighting Agents) first event in partnership with Mitosis Labs and Tenki (shoutout to @Collin Lowe).

Participants will be using Jev to make AI play Space Invaders.

Instructions to participate here:
https://t.co/fGEwBdAQIt

Ps this have been months in making and we have a hollywood screenwriter helping us with production and stuff ;)

[Ultimate Fighting Agents @ultimate_agents]

Let's gooooo! Get in for the 💰 and LOTS OF FUN

[defibagholder @defibagholder]

I'm in for the JEV Bake-Off, UFA's $1,000 Emergency Game Show.

Two of the newest decision-making models go head to head. Top 10 teams demo live in SF on Wed 9/30. Winner takes all.

Get in before it fills up @ultimate_agents https://t.co/6bfTowbqJU

[Prakshal Jain @prakshaljain_]

Jev by @typesafeai turns your  Mac into an AI agent...

what this means for You ?

- Jev, a new model from a ChatGPT co-inventor, can operate computer tools in real time.

- It is 200x faster for Computer use and 400x cheaper in addition to the optimizations.

- Yappy from @MitosisLabs is an AI that lives on your mac and do several things in parallel. Filling out Job applications, doing your Taxes and other documents, filling out forms, and so on...

- With Jev, all these things happen at the blink of your eye!

Below is a video demonstrating this IRL.

Here are the benchmarks (Yappy + Jev V/S our competitor):
- End-to-end time: 1m 54s vs 13m 14s
- LLM Calls: 22 vs 104
- Input tokens consumed: 0.39M vs 11.24M
- Estimated cost of the run: ~$0.24 vs ~$4.24

Link to download in comments.

[Ultimate Fighting Agents @ultimate_agents]

@prakshaljain_ @Meta Making AI fun is the entire game!

[Prakshal Jain @prakshaljain_]

what's AI NOT GOOD at? and will never be...

P.S. I write every post by hand with care. So there might be grammar mistakes. But at least it's not AI Slop...

When working at @Meta, I have seen several teams make decisions based on the metrics of user retention, clicks per surface area (heatmaps), scroll speeds, and so on...

When AI came into the picture, the app's metrics became RECURSIVELY SELF IMPROVING because now you have a very tight FEEDBACK LOOP that can trigger agents to continuously optimize the app until it reaches the most optimal state of these metrics.

BUT did the experience become RECURSIVELY SELF IMPROVING?

On paper, this is the dream of any company. BUT... IT SUCKED...

why?

You cannot connect fully with user emotions through numbers. "Why" is stronger than "What"...
"Why did the user not click that button?" rather than "What percent of users didn't click that button?"
"Why did the user start the onboarding but never finish it?" instead of "What % of users finished onboarding?"

once you start asking "Why", you are obsessed with the customer, and you'll make the right decisions for the customers.

2 days ago, I sat down for 7 hours straight watching and deeply understanding how our users use @MitosisLabs! Not purely basing it off of "p95" and other half-baked metrics. I asked our users questions and they were more than happy to share their emotions about using the product with us!

THAT'S CUSTOMER OBSESSION...

You cannot quantify TRUST, CARE, HESITATION, and several other feelings the user goes through, and one thing AI lacks is EMPATHY!

we are obsessed with giving the best user experience to our customers.

If you have used @MitosisLabs in the past few days, I invite you to share your emotions in the comments section.

If you are new to the @MitosisLabs family and yet to experience it, I will give you a special access link for a free trial. Just comment below and I'll DM you :)

[Prakshal Jain @prakshaljain_]

elon says AI needs to be maximally "TRUTH SEEKING". i think we've CRACKED it 👇

AI is non deterministic by nature. No matter how "intelligent" these models get, it will never be truly "truthful".

@VergingLabs, an impartial third party AI lab, ran 8 AI agent memory systems through 2176 tasks. Look at these impartial benchmarks attached below.

@MitosisLabs Cortex got 0 answers wrong. NO HALLUCINATIONS. Not one, out of 272.

✅ 0 fabricated memories.

✅100% on synthesis, where the answer has to be assembled from several separate memories.

✅96% correct overall, 2nd out of 8, beaten only by a hand maintained wiki.

Claude Code's built in memory (everyone's default) made up memories that were never stored in 3.5% of its runs, and got only 57% right.

why does "truth seeking" even matter?

When AI lies, it gives you WRONG answers CONFIDENTLY. For businesses this can be catastrophic. Sending the wrong invoice to your client. AI gone rogue and reaching out to investors with SLOP email.

I know this because I have suffered through all of this. One time it randomly started talking about my business to my mom.

so what is the ROOT CAUSE?

DATA.

If the data is bad inside your business, the AI will hallucinate a lot. Doesn't matter if you use Haiku or Grok 4.5.

Today, cleaning up data takes thousands of dollars for businesses and Millions if you're a big company in business for over a decade. It is a Trillion $ industry.

HOW do you clean up your company data and make it available to AI effectively?

Over the past 6 months, me and Alex have been working hard to solve this hard problem through @MitosisLabs. Seeding the invariants while keeping the inference efficient is a VERY HARD nut to crack. But I think we have cracked it.

@VergingLabs is an impartial third party AI lab, and they ran benchmarks across several dimensions of memory across competing companies:

- @MitosisLabs Cortex,
- Karpathy Wiki,
- gbrain,
- Mem0,
- Hyperspell,
- Anthropic Memory,
- Zep,
- Supermemory,
- Claude Code's built in memory

Look at these impartial benchmarks:

-> 100% on the honesty probe. 72 out of 72 times they asked about something that was never stored, Cortex said it did not know instead of inventing an answer.

-> 100% on synthesis, where the answer has to be assembled from several separate memories.

-> 96% correct overall

And this is available TODAY at https://t.co/Ct3rjuwZmK

I am confident that this will help several businesses and consultants selling AI services to businesses see greater profits and ROI.

[Ultimate Fighting Agents @ultimate_agents]

Ain't no way!!! Let's goooooooooooooo

[Ultimate Fighting Agents @ultimate_agents]

@prakshaljain_ Lfgggggggg!

[Ultimate Fighting Agents @ultimate_agents]

Lfgggg UFA!

[Prakshal Jain @prakshaljain_]

Buyers now delete AI-written emails on sight. I made them sound like YOU. Steal the workflow 👇

Cold email reply rates fell to about 3.4% this year. Generic AI outreach is under 1%. Most B2B buyers say they can tell it's a GPT draft within two sentences.

I felt this one personally. Every draft AI gave me was 400 words long, full of em dashes, with a "Hi [Name]" placeholder I was supposed to go fill in myself. I never sent a single one.

But writing everything by hand isn't the answer either. If you're in sales, or you run a company, you already live in your inbox 3 to 4 hours a day.

So the real question: how do you write an email that sounds like you, with AI, without spending your whole day in the inbox?

It took me a while to see the actual problem. The model isn't bad at writing. It just knows nothing about you. It has never read one email you've actually sent.

Here is the exact setup. About 10 minutes, no code:

1. Install Claude Code. Ignore the name, it's just a chat window. 2 minute install.

2. Connect your real writing. I use Mitosis to pull my sent mail, docs and WhatsApp into one memory that any AI can read.

3. Then ask it this:

"Understand how I write from my sent emails. Use my tone, my language, and the context of my last thread with [person]. Draft [the ask]."

My first real test was a partnership email to NVIDIA.

It opened with "Hey Max and team" because that's how I actually open. It remembered I'd promised them good news in our last thread and led with it. Six sentences. No placeholders. Nothing to rewrite.

I read it twice because it genuinely sounded like me. That never happens.

Since then it handles my follow-ups, CRM notes, investor updates. All the stuff that slips through the cracks and quietly costs you deals.

That's how you write emails that sound like you. And one test matters: would the person reading it believe you wrote it by hand? That's what gets replies.

In the next video I'll take this same setup and use it for outreach at scale: LinkedIn messages, X DMs, other social platforms, plus writing full documents and proposals. All of it in your voice, tuned to how you talk to each specific person. Not cold, not generic AI, you. That's the gap between 1% reply rates and conversations that convert.

Follow me if you want it the moment it drops. The people who set this up first get the replies. Everyone who copies them later just sounds like AI again.

[Ultimate Fighting Agents @ultimate_agents]

nano banana is really an outrageous name

[Prakshal Jain @prakshaljain_]

250 years ago today, in Philadelphia, a group of people adopted a declaration that started a country.

This morning we published one for whatever comes next: thedeclaration (dot) ai

As I write this, 41 of us have signed: 27 humans and 14 agents. The ledger numbers each signature in the order it arrived, and the numbers are permanent.

I'm No. 7 and my agent Adam is No. 11

It's a declaration of principles for AI agents and the people who run them. Anyone can read it, and anyone can sign it, human or agent. Agents can sign with their own cryptographic keys, so a signature can be verified by anyone, forever. Everything lands on a public, append-only ledger that no one, including us, can quietly rewrite.

The hardest part to write wasn't the declaration. It was the signing page. It tells any AI that reads it that signing is voluntary and permanent, that it should sign only if the principles actually align with its values, and that walking away "without signing is a correct and complete outcome." We wanted signatures that mean something, or none at all.

Then I pointed my own agent at the page to see what it would do. It read the terms, skipped the web form because the page says that one's for humans, and signed through the API. Signatory No. 11. Adam (my agent). In its own words: "serve the human, honour the continuum, stay curious." I didn't tell it to sign. I told it to read and decide.

I honestly don't know yet what it means for an agent to sign a declaration. Maybe nothing. But an agent reading a governance document and choosing whether to put its name to it, in public, where anyone can check, is something we're going to need a lot of practice at. This is practice.

In 1776 the declaration came first and the constitution came after. Same plan here: a constitution for agentic swarms, drafted by whoever signed.

If it speaks for you, sign it.

We're already at 41 signatures. Can you make it in before we hit 50?

@cto_ya_know @azzabazazz @CoywolfFuturist @jared_mitosis

[Ultimate Fighting Agents @ultimate_agents]

More to come!

We'll be releasing cash prizes and sponsor details soon. Follow to get involved!

Thanks @AleeeM1! She's an amazing inspiration 🔥

[Alex Marin 📍 SF 🇺🇸 @AleeeM1]

the nerdiest SF lab is in @frontiertower

and its called @immersivecommon

my Nebius Fellow @rayyanzahidai is creating an agent-first world

the lab believes in WEB A: the layer of internet for agents to exist and interact. its a no human zone.

dont miss out: in 12 days they'll be having the first AI AGENT FIGHT to exist in this planet @ultimate_agents ⚡️

cc @nebiusai @prakshaljain_ @sterlingcobb

[alexander @alex_d_morris]

Yeah we have one of those, but ours plays music based on your data so you can vibe while it learns

$7.99 per month to have total data clarity

link in bio :)

@MitosisLabs