SEGUNDA RONDA DE CAMBIOS Los límites del uso de razonamiento (GPT-5 thinking) van a subir. Eventualmente llegará a 3000 usos por semana, lo cuál es bastante más de los 200 usos actuales. Bien
MARKET TRENDS
-
AI Rollout Strategy Targets Mass Market Over Power Users
By
–
Pretty clear the rollout was not aimed at power users, but at solving a different problem. For better or worse.
-
GPT-5 Success Driven by Casual Users Adopting Reasoning Models
By
–
The numbers support my thesis: the biggest beneficiaries of GPT-5 are the casual users who will try out a reasoning model for the first time in their lives. The reasoning share in the free tier has risen from 1% to 7% and even to 24% for plus. Therefore, GPT-5 will be a success
-
Respect for xAI’s Rapid Development and Native Modality Progress
By
–
I don't know of any other company that has caught up and worked its way to the top in such a short time. Regardless of which model you like best, I have huge respect for @xAI 's performance and ongoing speed of development. Native modality incoming!
-
GPT-4o-mini Token Usage Spike: Large User Activity Theory
By
–
If you go back further you can see that gpt-4o-mini tokens went up for a couple of months and then went down, my guess is that it was just someone with big usage milking gpt-4o-mini. Doubt this actually means anything
-
O3 Adoption Low: GPT-5 Reasoning Will Surprise Users
By
–
Good to have numbers. And, and as suspected, almost nobody was using o3 before, even if they paid for it. For most users, experiencing a Reasoner will be a surprise expansion of AI’s abilities & GPT-5 will appear to be a very big leap when auto-routes to the Thinking version.
-
VC Perspective on Robotics Article by Terran Mott
By
–
Interesting article on robots from VC perspective by @TerranMott in @joinColossus magazine, however
-
ChatGPT users adapt to model changes despite preferences
By
–
Not so sure, people use ChatGPT regardless. They are just forced to use 5 instead of 4o
-
Compute shortage acknowledged, improvements expected soon
By
–
They've acknowledged that they're struggling with finding enough compute to run everything. I expect it'll get much better in the days ahead
