Simpler AI

Building the future of credit for micro-small businesses in South East Asia

[Simpler AI @simpler_ai]

Appreciate the traction we have been getting

The core premise is that the redaction industry was genuinely not that big PRE AI

But now POST AI, the industries that should use it the most like legal, healthcare, and finance are barred from doing so due to confidentiality red tape

By open sourcing our redaction engine, we hope that smaller firms for lawyers, doctors, and finance managers can utilise frontier intelligence safely and give a better deal to their customers

Be it better advisory or cheaper rates, we estimate between 12-16x time savings on traditional document redaction

We are working on an out of box experience for non technical users as well as variants specifically for legal, medical and financial industries with specific engine tuning

Thanks and much love to the builders who makes AI application great for those that really need it

Please feel free to take a look at our fully local offline SLM document redaction repo: https://t.co/yrWFUGNbK6

[Cloudflare @Cloudflare]

Agents need more than just a container to scale. We're introducing @cloudflare/computer, an agent runtime that dynamically orchestrates between fast, efficient isolates and full Linux containers to give every agent a computer of its own. https://t.co/I708jAY59z

[Simpler AI @simpler_ai]

point proven for what we are striving towards

[a16z @a16z]

.@DecagonAI co-founder and CTO Ashwin Sreenivas says the choice between expensive frontier models and cheaper, less capable ones is a false trade-off:

"Even if you have a 'dumber model,' you can get it to higher performance on that specific task."

"When we fine-tune smaller, dumber models, it's that they're just not as general purpose, but on the specific task we want them to do, they actually outperform the large, smart, state-of-the-art models."

"So we end up getting all three things. It is better at the task, it is cheaper, and it is faster."

@AshwinSreenivas

[Simpler AI @simpler_ai]

@santiagomed if you have the tokens to spare, get the model to show it to you as a visual

[Simpler AI @simpler_ai]

This is huge, looking forward to it

[Matthew Prince 🌥 @eastdakota]

Buries the headline: “We’ll open source next week!” #staytuned

[Simpler AI @simpler_ai]

most decisions are just two-way doors

if you can walk back through, just go. 70% of the info is usually enough

waiting for 90% means you’re already slow

80% correct and moving > 100% correct and stuck

[Simpler AI @simpler_ai]

Prompt engineering is dead.

You don’t need clever word tricks anymore.

The real skill now is goal alignment: clearly telling AI what success looks like and why it matters, so it actually aims for the outcome you want, not just the exact words you typed.

Stop polishing prompts. Start defining the goal.

[Simpler AI @simpler_ai]

@thdxr i mean if it works it works

[Simpler AI @simpler_ai]

it would appear that x402 is the new meta in this round of AI verticals

[Revolut @Revolut]

🤝

[Jett @jontangit]

chemical x to be power puff founders

[Simpler AI @simpler_ai]

@andrewchen what do you think will happen after we reach peak intelligence in models? will the floor raise across the board as the next priority?

[Simpler AI @simpler_ai]

@Thera_drop probably the latter

[Simpler AI @simpler_ai]

@theo didnt you permanently cancelled your claude sub?

[Simpler AI @simpler_ai]

south korea is prosperous but doing so well is an over statement, the distribution of wealth is extremely uneven, the permanent underclass is a reality over there which leads to over leveraging on what should traditionally be something that doesnt swing so much (stock market index)

[Simpler AI @simpler_ai]

personal favourite of mine, downside protection first!

[WOLF @WOLF_Financial]

CHARLIE MUNGER'S JOB IN WORLD WAR II WAS CLEARING PILOTS TO FLY. SO HE ASKED HOW HE WOULD KILL THEM

Munger was an Army meteorologist, drawing weather maps and approving flights. His favorite trick, "the inversion process," started there.

"Suppose I wanted to kill a lot of pilots." He worked out there were only two easy ways: icing the planes could not handle, and conditions that left a pilot without enough fuel to land safely.

Then he spent the war keeping every flight he cleared miles away from both.

Decades later he ran the same trick on everything. Asked how he would fix a country, he said he would first list what would most easily hurt it, then avoid all of it.

"It's the same thing, it's just in reverse." Most people ask how to succeed. Munger asked how to fail, then stayed away from the answer.

[Simpler AI @simpler_ai]

Asia has a massive credit problem mainly between 21-40, where blended you are looking at maybe 85%~ of population on some form of credit?

As home prices increase and wages not so much, many younger adults feel that owning a house is near impossible unless they take massive swings which is where all of the over leveraging comes in

it doesnt help that south korea specifically has a materialistic and looksmaxxing culture due to kpop which further increases credit spend to fit in

[Simpler AI @simpler_ai]

@0xPia in progress

[Simpler AI @simpler_ai]

not competing in early stage investing i agree with, you give the benefit of some kind of traction and settle for less due to these premiums

however as im a builder now, i too agree from business fundamentals

perfect competition only brings the variable of price down and then youre competing on who has a better price per metric of value

instead focus on the fundamentals of being so differentiated that no rival can offer a close subsitute

invest for the long term and capture the value being created

[sam lessin 🏴‍☠️ @lessin]

Never Compete. I was on with someone from our team at slow talking about competition recently… ‘how do you compete in venture capital’, etc… win deals. My honest philosophy?

I hate competition professionally (love it in sport, but not business)…. And I think competition is especially stupid in early stage venture capital, where you are competing over / crawling over each other for out of the money call options on things that almost certainly will not work.

Early stage VC competition on deals almost by definition leads to over-pricing / overpaying … and it also demonstrates a fundamental lack of creativity. Egotistically, if you want your money to matter / to make a difference, you want to be funding otherwise unfunded opportunities — where you see something others don’t… where money is most expensive, commands the highest return AND - to put a nice spin on it - where the money matters the most.

So if you find yourself in a competition over a seed deal, IMHO you are a bad capitalist / limiting your returns, an un-creative person / a mere ‘market participant’ (yuck), and also misusing the incredible mandate and license you have to find things and make them happen in the world.

You have to be careful about seeing mirages, because some things / most things are not funded because they are actually bad.. you can’t like things just because they are unliked… BUT your job is to have discipline in the wilderness / in the wander, and as the game evolves … both as a capitalist, and egotistically as someone who realizes how awful it is for your tombstone to say ‘market participant’

[Simpler AI @simpler_ai]

@lessin ahh one of the classics "competition is for losers" by Thiel

[Simpler AI @simpler_ai]

@seema_amble can relate

[Simpler AI @simpler_ai]

All 3 can be true to varying degree

Preliminary testing suggests that stronger software rails specifically for small local models are enough to get most things done

But interestingly on frontier models, they are not bottlenecked by the rails but instead flow faster, to the point it confidently make mistakes

So smarter models needs more checkpoints than dumber models