Artificial intelligence is officially putting architects out of work. Now creating impressive architectural designs is even simpler than using ChatGPT. I showed you how to do it:
GENERATIVE AI
-

DeepSeek R1 Gains Self-Correction and Creative Worldbuilding Capabilities
By
–
Le nouveau DeepSeek r1 est en fait plutôt bon. Il est désormais capable de corriger sa propre chaîne de pensée comme o3, et de faire du worldbuilding créatif comme Claude.
R1 n’en était pas capable auparavant. Ils vont bientôt publier les poids du modèle, pour que le monde -
ChatGPT estimates 500 hours for invoice processing automation
By
–
ChatGPT, happy to speculate on this, produced a model which suggested it is probably ~500 hours or one solid quarter of full professional effort doing nothing but invoices. That feels approximately right to me. (I do more of this than most people; ran a business since 2006, etc)
-
LLMs Enable PDF Parsing for Financial Applications
By
–
(One reason I think nobody seems to have tried implementing this prior to LLMs is that parsing arbitrary PDFs well enough to trust money to the results was almost intractably hard, and now it's something which is probably single digit engineers times single digit weeks to build.)
-
LLMs Automating Redundant PDF Data Entry Tasks
By
–
I do not even want to speculate how much of my life I've spent parsing PDFs to redundantly type payment instructions into a web application. I will not regret that portion of the job getting eaten by LLMs and good experiences built on top of them.
-
Mercury’s AI-Powered Send Money Feature Parses PDFs
By
–
A tiny product experience which was just great that I have to share: Mercury's new Send Money feature lets you either do it the usual way or upload e.g. a bill. I uploaded a capital call document. An LLM (presumably) parses the PDF and prefills everything, displaying inline…
-
Leonardo AI Motion 2.0 Feature with Custom LoRA
By
–
Created in @leonardoai with a custom LoRA and then used their new Motion 2.0 feature.
-

DeepSeek-R1-0528 Now Available on Hugging Face
By
–
DeepSeek-R1-0528 is now on Hugging Face! https://
huggingface.co/deepseek-ai/De
epSeek-R1-0528/tree/main
… -

Large Reasoning Models Self-Training Capabilities Study
By
–
Can Large Reasoning Models Self-Train?
Paper: https://
arxiv.org/pdf/2505.21444
v1.pdf
… -

Self-Consistency Training for LLM Scaling Without Human Supervision
By
–
Scaling LLMs through Reinforcement Learning (RL) usually needs human-crafted verifiers or gold answers, which limits scalability. Can models train themselves, without external supervision? This paper propose: using the model’s own self-consistency (i.e., agreement across
