Articles by maxutility
10

Google has a secret reference desk (cardcatalogforlife.substack.com)

1

The decade long feud shaping the future of AI (wsj.com)

3

Refine: AI-Powered Peer Review (refine.ink)

1

Politics and the English Language, January 2026 edition (daringfireball.net)

2

The most popular blogs of Hacker News in 2025 (refactoringenglish.com)

3

The 26 Most Important Ideas for 2026 (derekthompson.org)

2

The Problem with Letting AI Do the Grunt Work (theatlantic.com)

2

AI is mastering language. Should we trust what it says? (2022) (nytimes.com)

2

Heiliger Dankgesang: Reflections on Claude Opus 4.5 (hyperdimensional.co)

2

Andrej Karpathy on improved Tesla FSD performance under HW4 (twitter.com/karpathy)

2

How the Earth Invented Sleep (derekthompson.org)

3

Did the ChatGPT Erdos controversy obscure a real achievement? (twitter.com/sebastienbubeck)

2

Will Any Crap Cause Emergent Misalignment? (lesswrong.com)

1

Mantic: A New Kind of Foresight (mantic.com)

1

A better way to think about AI (theatlantic.com)

5

The Substack Branding and Faux Prestige Trap (daringfireball.net)

2

Ask HN: How is Generative AI being used in AAA game development today?

1

AI and the Last Mile: Discerning human judgement as a luxury good (hollisrobbinsanecdotal.substack.com)

1

Rage bait is all the rage on Threads – so I tried it (businessinsider.com)

1

What to do with unwanted political spam texts (daringfireball.net)

2

Scientists have created a nutritious powder from plastic-consuming microbes (undark.org)

3

Silicon Valley's Trillion Dollar Leap of Faith (theatlantic.com)

1

All the Storytelling Frameworks Together (twitter.com/acagamic)

12

GPT-4o's Chinese token-training data is polluted by spam and porn websites (technologyreview.com)

1

Netflix and A24 both land in hot water over apparent AI stuff (avclub.com)

1

AI Has a Measurement Problem (nytimes.com)

2

Ask HN: Could predicting software output be used for synthetic data?

2

America's Magical Thinking about Housing (theatlantic.com)

1

Cerebrum 7B – Mistral fine-tune created specifically for reasoning tasks (huggingface.co)

27

Claude 3 System Prompt (twitter.com/amandaaskell)

2

Apple Reinstates Epic's EU Developer Account (daringfireball.net)

5

Altman says OpenAI investigation will soon close, staff brace for more surprises (businessinsider.com)

6

Wyoming hits the rare-earth mother lode (wsj.com)

1

Bluesky Is Back (platformer.news)

5

What the Japan and Alaska Airlines Incidents Tell Us about Airline Safety (nytimes.com)

2

The on-device AI era is closer than it seems (supervised.news)

4

Why Substack Is at a Crossroads (platformer.news)

4

Stolen Checks Are for Sale Online. We Called Some of the Victims (nytimes.com)

3

The AGs reveal their hand against Meta (platformer.news)

1

OpenAI's Alignment Problem (platformer.news)

126

Inside The Chaos at OpenAI (theatlantic.com)

4

The plane tech that could make turbulence a thing of the past (thetimes.co.uk)

2

A robot's nighmare is a burrito full of guac (theatlantic.com)

5

iOS 17 forces woman named 'Siri' to change her name (cultofmac.com)

3

Don't Be a Decel(erationist) (noahpinion.blog)

3

Chipotle robots may soon construct your salads and bowls (washingtonpost.com)

1

Ask HN: Unintended Consequences of Ambient Pressure Room Temp. Superconductor

3

What’s the Deal with Sensor Tower? (daringfireball.net)

46

Hearing aids may cut risk of cognitive decline by nearly half (washingtonpost.com)

10

Has Covid’s Patient Zero Finally Been Named? (theatlantic.com)

6

The Government Must Say What It Knows About Covid’s Origins (nytimes.com)

4

FTC Sues Amazon for Tricking Users into Subscribing to Prime (nytimes.com)

2

Car-Rental Companies Are Ruining EVs (theatlantic.com)

3

Ask HN: Secure way to send/access a document to/on a public computer?

6

Elon Musk has picked new CEO for Twitter (washingtonpost.com)

2

Tell HN: Unreported GPT-4 outage for at least 24 hours

2

The Inferno of the Nerds (aiascendant.substack.com)

2

Ask Marvin – AI Functions in Python (askmarvin.ai)

2

The biggest problem in media is the audience (slowboring.com)

1

The Kottke.org Ask Me Anything (kottke.org)

4

Cuboid skyscraper “big enough to hold 20 Empire State Buildings” (newatlas.com)

3

The AI Disaster Scenario (theatlantic.com)

18

Lab leak is not a debunked conspiracy theory and never has been (slowboring.com)

4

Jason Kottke: Ask Me Anything (kottke.org)

2

What’s our Problem? A (Wait but Why?) self-help book for societies (waitbutwhy.com)

1

What’s Our Problem? A (Wait but Why?) self-help book for societies

3

The Death of the Smart Shopper (theatlantic.com)

1

H3: New generative language model that can outperform Transformers (twitter.com/realdanfu)

1

Psychedelics open your brain. You might not like what falls in (theatlantic.com)

43

Cistercian Numerals (kottke.org)

1

Google’s new AI turns text into music (theverge.com)

4

The billions dollar VR/AR headset question (daringfireball.net)

5

TikTok’s Enshittification (pluralistic.net)

2

AI is improving faster than most humans realize (bloomberg.com)

3

Alarmed by A.I. Chatbots, Universities Start Revamping How They Teach (nytimes.com)

1

Did a Fourth Grader Write This? Or the New Chatbot? (nytimes.com)

2

Consolidation-disruption index: Science has a crummy paper problem (theatlantic.com)

2

I tried to learn as much Deep Learning math as I could in 24 hours (lesswrong.com)

2

Transformers from Scratch (2021) (e2eml.school)

5

zxcvbn: Public Low-Budget Password Strength Estimation (2017) (github.com/dropbox)

2

Illustrated Guide to Transformers Neural Network: A step by step explanation (youtube.com)

4

How AI Image Generators Work (youtube.com)

5

Tell HN: ChatGPT Writes an Electrician’s Manual

4

Mario,LeBron,Trump: Twitter is playing whack-a-mole with ‘verified’ users (theverge.com)

9

Elon Musk is bad at this (theatlantic.com)

4

Facebook's Future: My Sad, Lonely, Expensive Adventures in Zuckerberg’s V.R (nytimes.com)

2

I won my three-year AI progress bet in three months (astralcodexten.substack.com)