Top all-time · #2598
Alan D. Thompson
Personal blog · lifearchitect.ai · on HN as adtai
Alan D. Thompson writes LifeArchitect.ai, a personal blog at lifearchitect.ai about ai. Since 2021, 29 different Hacker News readers have linked it 136 times, most often “33TB of text data for a 1T-parameter model”. The last post was in 2014, but the archive is still widely read.
When Microsoft writesa paper on AI,they go to Alan. LifeArchitect.ai is the 'gold standard' for understanding, visualizing, and realizing the capabilities of post-2020 AI. With over 100 papers and articles, 300 videos, and regular editions of The Memo, LifeArchitect.ai provides insights to all major
Most-linked posts?
- /whats-in-my-ai/4 people · 6 mentions · first linked 2023-02-15
- 33TB of text data for a 1T-parameter model4 people · 4 mentions · first linked 2023-02-04
- The gap between open source and corporate LLM sizes is increasingly widening3 people · 5 mentions · first linked 2022-07-22
- Alan's conservative countdown to AGI: 99%3 people · 3 mentions · first linked 2024-06-21
- Inside Language Models (from GPT-3 to PaLM)3 people · 3 mentions · first linked 2022-12-04
- GPT-3.5 IQ testing using Raven’s Progressive Matrices3 people · 3 mentions · first linked 2022-12-26
- /models-table/2 people · 52 mentions · first linked 2023-05-11
- Alan's ASI Checklist2 people · 5 mentions · first linked 2025-11-05
- /memo/2 people · 4 mentions · first linked 2023-04-07
- /gpt-4/2 people · 3 mentions · first linked 2023-03-17
What Hacker News says?
Since OpenAI has gone completely closed, they've been increasingly opaque and dodgy about how even things like basic chat works. Assuming the various leaked details of GPT-4 [1] are correct (and to my knowledge there has been no indication that they are not), they have been actively misleading and deceptive - as even the 'basic' GPT4 is a mixture of experts system, and not one behemoth neural network. [1] - https://…
Comment linking /gpt-4/
on “GPT-4o”Submitted The New Irrelevance of Intelligence as a story · 17 points
HN discussionDefinitely not. Here's 300+ LLMs without derivatives: https://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Meta Llama 3”Submitted O3 Evals as a story · 7 points
HN discussionhttps://lifearchitect.ai/agi/
Comment linking /agi/
on “Safe Superintelligence Inc.”Submitted GPT-3 would have easily beaten IBM Watson on Jeopardy! (probably) as a story · 6 points
HN discussionI think Alan's conservative countdown to AGI will need to be updated after this. https://lifearchitect.ai/agi/ This is really impressive stuff. I thought about it a couple of months ago, that probably this is the next modality worth exploring for data, but didn't imagine it would come so fast. On the other side, the amount of compute required is crazy.
Comment linking /agi/
on “Diffusion models are real-time game engines”Submitted 33TB of text data for a 1T-parameter model as a story · 3 points
HN discussionhttps://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Claude Opus 4.5”Submitted Alan's conservative countdown to AGI: 99% as a story · 1 points
HN discussionThe EuroLLM-9B model release is from Dec/2024, and scores just above random chance for benchmarks like MMLU-Pro (17.6%, random chance is 10%). Comparison with similar EU models + 600 other highlights: https://lifearchitect.ai/models-table/
Comment linking /models-table/
on “EuroLLM: LLM made in Europe built to support all 24 official EU languages”Submitted Pre-ASI Society as a story · 1 points
HN discussionHere's something that talks about it. I can't speak for the legitimacy, but I'm not pulling it out of my ass. They may be pulling it out of theirs. :-) https://lifearchitect.ai/gpt-6/
Comment linking /gpt-6/
on “Claude 3.5 Sonnet”Submitted Alan's ASI Checklist as a story · 1 points
HN discussionOld paper (Oct/2023), but the weights are new (Dec/2023): https://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Ferret: A Multimodal Large Language Model”Submitted GPT-6 (2025) as a story · 1 points
HN discussionhttps://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Mistral Small 3”Submitted The gap between open source and corporate LLM sizes is increasingly widening as a story · 1 points
HN discussionhttps://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Ember-1”Submitted GPT-3 in Google Sheets as a story · 1 points
HN discussionThought this was going to be about Dr Alan D. Thompson's Leta AI! https://www.youtube.com/playlist?list=PLqJbCeNOfEK88QyAkBe-U... https://lifearchitect.ai/leta
Comment linking /leta
on “Social media influencer/model created from AI lands 100 sponsorships”Submitted LLM scaling laws, in plain English as a story · 1 points
HN discussionModel card: https://github.com/meta-llama/llama-models/blob/main/models/... On the Models Table: https://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Llama-3.3-70B-Instruct”Submitted Inside Language Models (from GPT-3 to PaLM) as a story · 1 points
HN discussionThat's 3 releases for Mistral in 24 hours. https://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Mistral NeMo”Submitted GPT-3.5 IQ testing using Raven’s Progressive Matrices as a story · 1 points
HN discussionhttps://lifearchitect.ai/models-table/
Comment linking /models-table/
on “Eagle 7B: Soaring past Transformers”Submitted Integrated AI: The sky is bigger than we imagine (mid-2022 AI retrospective) as a story · 1 points
HN discussionRobots are only harder because they have expensive hardware. We already have robots that can load dishwashers and do other manual work but humans are cheaper so there isn't much of a market for them. The rising tide idea came from a 1997 paper by Moravec. Here's a nice graphic and subsequent history https://lifearchitect.ai/flood/ Interestingly, Moravec also stated: "When the highest peaks are covered, there will be…
Comment linking /flood/
on “The Bitter Lesson Is Misunderstood”Param estimates etc: https://lifearchitect.ai/olympus/
Comment linking /olympus/
on “Amazon Nova”
