Sabitlenmiş Tweet
FP Marcil
7.7K posts
The main metric that matters is price per task completion.
Second is speed to task completion.
Nothing else matters.
English
Building a spammy site that gets indexed in search and picked up/cited by a language model - before later dropping out of the search results and not being cited for anything relevant in AI search a couple weeks later - is not a strategy.
English
@MatthewBerman "The more likely scenario is OpenAI/Anthropic agents are just able to do…everything…one day soon"
Highly unlikely, and nobody would want models to do everything anyway.
English
As the person who mainstreamed the “SaaS is Dead” idea, the vast majority of companies will not vibe code a CRM.
There’s too much risk in getting something wrong. Too much maintenance. Ultimately companies need someone else to blame when software breaks.
The more likely scenario is OpenAI/Anthropic agents are just able to do…everything…one day soon.
Harry Stebbings@HarryStebbings
"We replaced Salesforce with a vibe-coded CRM built for our own workflows. The custom system integrated our AI agents more effectively, worked better for the team, and made Salesforce unnecessary. That decision cut a $600,000 annual software bill to zero." Is this an anomaly or the start of a much larger trend @chamath @Avishai_ab @jasonlk @benioff
English
Unpopular opinion: AI basically became useful for enterprise use early this year with agentic AI acceleration. Everything before was basically consumer and hobbyist. 2026 is year 0 for AI enterprise use. So why would anyone have expected measurable productivity increase already.
English
If Intelligence = inventive force instead of applicative knowledge, we've made no progress. It's not exponential growth; progress in intelligence is on a slight linear path, and the harnesses might have more to do with making it useful for something than with the models themselves at this point.
This graph should be transformed to show the human in the middle, a circle where everybody is potentially as expert as each other (with practical experience/skill divergence), and we all have at our fingertips the work of 3-4 generations who created the knowledge economy. Current AI is a replacement for the information in universities and a way to use this knowledge much faster. This is exciting enough, but it's not a significant increase in the machine's intelligence. We would feel that compound much faster.
The small "discoveries" of LLMs seem to me like a demonstration of solving the scarcity of logic work on problems we already know about (Math). It's a type of discovery that requires no inventive power; rather, it's one where the human is data-limited. How many other problems are unsolvable by humans because we are data-limited (we can't hold enough information)? Exciting.
Ants are still far more intelligent than LLMs, particularly when working in groups.
English
@CliffordSosin Do you have evidence 'that the explosion never came', and that we aren't on this path pictured?
English
@boxmining Or, we're done with growth hacking; here's the new reality.
English
in other words:
Chinese competition is tough. so here is more Fable 5
Claude@claudeai
Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit. Demand for Fable has been challenging to predict, which is why we rolled it out to subscription plans in stages, extending access several times as we secured additional capacity.
English
@TimJayas @shiri_shh How can they surpass them if they are built against them?
English
@shiri_shh Chinese models are finally on the same level of US AI labs
And soon they will surpass them with lower pricing
English
Kimi K3 vs Fable 5
$16.45 vs $37.94
Kimi.ai@Kimi_Moonshot
Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for long-horizon agentic coding and self-evolving workflows Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Open Weights by July 27, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.com/blog/kimi-k3
Español
@AIandDesign Was it always the plan, maybe?
The time-limitation thing early was a growth hack for extra subs
English
We won, comrades. We won. We fucking won.
Claude@claudeai
Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit. Demand for Fable has been challenging to predict, which is why we rolled it out to subscription plans in stages, extending access several times as we secured additional capacity.
English
@DavidGQuaid who is ever going to solve the peanut butter jar automatic re-order app is going to make an absolute bomb
English
@FPMarcil I have to vibe code them - which means I have 3 apps to manage my fishbowl filled with shrimp...
And I haven't finished my Ant Colony Step Counter SaaS app yet.... trying to vibe code my own MDR/SIEM to keep it secure.....
English
Ok — I need to have a serious pep talk with all of you 🤨
It's time to unite the Marketing Web Community 🤝
🔍 SEOs
✍️ Content Writers
🎯 Brand Marketers
🛠️ SaaS Builders
It turns out some of you STILL haven't co-founded a GEO tool company with 6 of your besties 👯 and at least one co-founder in a remote part of the world 🌴 (I've heard Borneo is a great and unexplored resource for this)
What's going on guys?? We agreed on at least 2 GEO tools for every person actually using one 📊
Let's get up and go! 🚀
English
You are going to be able to run Fable 5 locally on your desk
In 2 years Apple will be coming out with Mac Studios with 1.5TB of memory
With just 300gb of memory you can run Opus 4.8 level intelligence
Think of what you can do with 5x that
You need to be preparing for this now
Start getting familiar with local AI technology
Go to your Hermes/OpenClaw and use this prompt:
“I am brand new to local AI and want to get familiar. Look at the computer you are currently on. Understand the specs. Then go on Huggingface and find the best models I can run on it. Then, walk me through how these models work, how they will run locally, and use cases I can do with them. After walking me through all of that so I’m educated, you can then load it onto this computer and build an interface so I can use them”
In 2 years EVERYONE on Earth will have a local model running on their desk
The people who start preparing now will be WAY ahead of everyone else
AppleTrack@appltrack
Apple's M7 Ultra chip coming in 2029 is rumored to support 1.5TB of RAM. This would make the processor much more capable for on-device AI.
English
Is there a mixed-location business model out there for inference? The 2 current choices: 1) data centers, with big model brands (high quality, high price 2) open-source options running on my machine. Why wouldn't Anthropic/OpenAI/Grok/Google/Meta sell distributed inference licenses (like software) of their model without showing their proprietary stuff? Still send all the data back to the mothership.
"We are compute constrained"
"100s million CPU/GPUs with 0% utilization in their user's computers"
Distributed licenses??
AI model companies can cut their data center build-up; energy is already online, etc. Revenue becomes more like software; marginal cost approaches zero.
What is the fear?
English
@VeryWellVersed If you didn't need QEMM, it didn't happen.
Just kidding.
English
Yes, I am except the numbers were never anywhere near this high.
Ramin Nasibov@RaminNasibov
are you this old?
English
Is Microsoft ditching patents ? Same issue.
You may think your IP is protected. It's not. Every model is training on every patent app, and simultaneously being prompted to find workarounds or ways to compete with that patent
I think the value of a patent has dropped and the value of a trade secret has increased
Vaibhav Sisinty@VaibhavSisinty
Satya Nadella just warned every company using AI: you are paying twice. Once with money. Again with something far more valuable. He published an article introducing something called the Reverse Information Paradox. And it changes how you think about every AI tool your company uses. Nobel laureate Kenneth Arrow described the original paradox: a seller risks giving away knowledge just to sell it. Nadella says AI flips this completely. In the AI age, the buyer gives away knowledge just to use what they bought. Every time your team uses Claude or GPT at work, every prompt reveals what you are building. Every correction teaches the model what good looks like inside your company. Every eval shows what you value. Every trace exposes your workflow. The model provider learns more about you with every interaction. You learn almost nothing about what they are learning in return. Your corrections are distilled institutional know-how. The kind a competitor could never buy. And it leaks trace by trace, correction by correction, without you noticing. His line: "You can offload a task. You can offload a job. But you can never offload your learning." If the model provider disappears tomorrow, do you still own the intelligence your team built on top of it? Your evals. Your memory. Your traces. Your workflows. Or did all of that compound inside someone else's infrastructure? In the cloud era, companies accumulated data. In the AI era, they accumulate learning. Right now, most of that learning is compounding inside the model provider. Not inside the company paying for it. The CEO pushing AI harder than anyone just told you to protect your knowledge from the very tools he is selling you. That should tell you everything.
English
@DaveShapi Or it's a growth hacking way to create a perception of scarcity and get their customers used to this level of pricing. Each time they release a new model, they can raise the profit margin.
English
Here's my honest inference:
1) Anthropic is genuinely a schizophrenic and paranoid institution. They honestly thought that releasing this model would destroy the world or something. So their first intention was to do a "test release" and integrate that feedback so they can start loosening the guardrails.
2) But then OpenAI released GPT 5.6, which is superior on UX and several benchmarks (plus it is not hobbled) and Anthropic saw a surge of cancellations.
3) So now they keep kicking the can down the road, probably a bit of a "split brain" reasoning of "but what about RSI" versus "we must mog OpenAI"
That's the intuition I have since I follow this whole thing much more than is healthy.
Chase Brower@ChaseBrowe32432
Is there anyone who can steelman the “tell your customers they’re losing their favorite product in 5 days, then say sike over and over” strategy
English