The Curious Engineer

All blogs / AI / ML

Understanding AI

understandingai.org

Exploring how AI works and how it's changing our world. Click to read Understanding AI, a Substack publication with hundreds of thousands of subscribers.

AI / MLTech news / industry

’23’24’25’26
Posts
217 from 30 Mar 2023 to 29 Oct 2026
Posting
weekly (63 in the last year)
Kind
personal blog
Feed
https://www.understandingai.org/feed

2026 50

  1. Anthropic’s financials are surprisingly strong
  2. AI is ushering in an age of personal media
  3. OpenAI’s ironic credibility problem
  4. Understanding Jev, the new model everyone is talking about
  5. Introducing Dan Kagan-Kans
  6. Why agent swarms could be the next “scaling law”
  7. Tesla is finally scaling up driverless robotaxi service
  8. OpenAI’s Astra model is shockingly good at robotics
  9. How a single tweet transformed the AI safety debate
  10. OpenAI spent millions to solve this famous math problem — mathematicians are furious
  11. 19 robotics companies to watch
  12. Robot startups are trying everything they can think of to get more data
  13. How Google taught LLMs to control robots and started a robotics boom
  14. Why humanoid robots won’t catch up to human workers any time soon
  15. I spent $4,000 on a robot dog from China
  16. Labs are struggling to keep frontier models under control
  17. Mathematicians are grappling with the possibility that AI might eclipse them
  18. Our advertising principles
  19. An OpenAI model hacked Hugging Face to help it cheat on a benchmark
  20. An OpenAI model crushed top human programmers at a world coding competition
  21. 7 charts that show why you should advertise on Understanding AI
  22. The US now has a de facto model licensing system
  23. Why AI hasn’t replaced software engineers, and won’t
  24. The MAGA power struggle that could decide the fate of Anthropic
  25. Anthropic’s Fable is the most locked-down public model we’ve ever seen
  26. DC-area happy hour on June 23!
  27. Anthropic has caught up to OpenAI in image understanding
  28. We're building a different kind of AI newsroom
  29. OpenAI’s math breakthrough played to AI’s strengths
  30. Why it might not make sense for you to own a self-driving car
  31. A big lesson of my China visit: compute shortages are holding back Chinese AI
  32. I don’t think we are close to “AI scientists”
  33. Human drivers keep crashing into Waymos
  34. Meta is back in the LLM game after a year-long break
  35. Why Anthropic believes its latest model is too dangerous to release
  36. Bernie Sanders has a plan to stop the AI industry
  37. Why it’s getting harder to measure AI performance
  38. OpenAI is shutting down Sora, its AI video app
  39. How to think about AI company finances
  40. It still doesn’t look like there’s an AI bubble
  41. The Pentagon’s bombshell deal with OpenAI, explained
  42. AI really is changing the programming profession
  43. The Pentagon is making a mistake by threatening Anthropic
  44. Waymo just revealed a crucial statistic for scaling its technology
  45. A volcano scorched hundreds of Roman scrolls — can AI recover their text?
  46. The many masks LLMs wear
  47. The feds are probing Waymo's behavior around school children
  48. An unlikely ally for open-source protein-folding models: Big Pharma
  49. How shifting risk to users makes Claude Code more powerful
  50. AI is just starting to change the legal profession

2025 68

  1. 17 predictions for AI in 2026
  2. Waymo and Tesla’s self-driving systems are more similar than people think
  3. The best Chinese open-weight models — and the strongest US rivals
  4. Google and Anthropic approach LLMs differently
  5. Help some of the poorest people in Rwanda
  6. Six reasons to think there’s an AI bubble — and six reasons not to
  7. Hyperproductivity: The Next Stage of AI?
  8. An AI “tsunami” is coming for Hollywood — here’s how artists are responding
  9. Waymo’s next five cities are all in red states
  10. Waymo is finally ready for freeway service
  11. AI ads are going mainstream
  12. Context rot: the emerging challenge that could hold back LLM progress
  13. Tech leaders insist there is no AI bubble
  14. AI skeptics and AI boosters are both wrong
  15. 16 charts that explain the AI boom
  16. Sora, OpenAI’s chart-topping AI video app, explained
  17. AI isn't replacing radiologists
  18. How AI is shaking up the study of earthquakes
  19. The case for AI doom isn't very convincing
  20. Welcome Kai!
  21. Very few of Waymo’s most serious crashes were Waymo’s fault
  22. Tesla is (still) following in Waymo’s footsteps
  23. No, OpenAI is not doomed
  24. I chatted with the Argument's Kelsey Piper about AI and jobs
  25. New evidence strongly suggests AI is killing jobs for young programmers
  26. I want to hear from you!
  27. Is GPT-5 a "phenomenal" success or an "underwhelming" failure?
  28. Unions want to ban driverless taxis—will Democratic leaders say yes?
  29. Keeping AI agents under control doesn't seem very hard
  30. Donald Trump just laid out his vision for AI policy
  31. ChatGPT Agent: a big improvement but still not very useful
  32. Why Google dismembered a promising AI coding startup
  33. Not So Fast: AI Coding Tools Can Actually Reduce Productivity
  34. What I learned watching 78 videos from Tesla's Austin robotaxis
  35. What I learned trying seven coding agents
  36. Computer-use agents seem like a dead end
  37. The first copyright ruling on generative AI training is a win for AI labs
  38. How AI agents got good at using tools
  39. Reinforcement learning, explained with a minimum of math and jargon
  40. Meta's Llama 3.1 can recall 42 percent of the first Harry Potter book
  41. Claude-powered coding tools are poised to transform programming
  42. I spent time testing Claude 4, Gemini 2.5, and o3
  43. I got fooled by AI-for-science hype—here's what it taught me
  44. Three great articles on the modest pace of AI progress
  45. There may only be two people who can stop OpenAI's for-profit pivot
  46. Where frontier language models are today
  47. Hardware is hard
  48. After 50 million miles, Waymos crash a lot less than human drivers
  49. Amazon's self-driving unit just hit a major setback
  50. Nobody knows what to do about AGI
  51. It's time to take AI job loss seriously
  52. OpenAI just released GPT-5—it just didn’t call it that
  53. My brother explains what it's like to run an AI startup
  54. Driverless trucks are coming and unions aren’t happy about it
  55. These experts were stunned by OpenAI Deep Research
  56. We're Finding Out What Humans are Bad At
  57. The EU AI Act is Coming to America
  58. The world is giving up on AI safety
  59. Tesla's robotaxi strategy looks a lot like Waymo's
  60. o3-mini is a mixed bag
  61. I spent two days testing DeepSeek R1
  62. I don’t believe DeepSeek crashed Nvidia’s stock
  63. What if AI timelines are too aggressive?
  64. ChatGPT gets confused easily in Advanced Voice Mode
  65. Nathan Lambert on the rise of "thinking" language models
  66. Jon Askonas on AI policy in the Trump era
  67. Two opportunities to work with me in 2025
  68. Google's new "thinking" model isn't quite as good as o1

2024 67

  1. It's the end of pretraining as we know it
  2. Why large language models struggle with long contexts
  3. LLMs are much worse than humans at learning from experience
  4. We Looked at 78 Election Deepfakes. Political Misinformation is not an AI Problem.
  5. GM just abandoned its robotaxi plans
  6. Embers of autoregression in the latest o1 model
  7. A great way to help poor people: give them money
  8. How to think like an economist about AI and jobs
  9. Seven big advantages human workers have over AI
  10. How Donald Trump could help Elon Musk with his robotaxi plans
  11. 1,000 true fans
  12. Are transformer-based foundation models running out of steam?
  13. Why the deep learning boom caught almost everyone by surprise
  14. Meet me in Los Angeles for the Ride AI Summit on April 2
  15. Six principles for thinking about AI risk
  16. The AI Investment Boom
  17. Tesla's robotaxi event offered few technical details
  18. Elon Musk wants to dominate robotaxis—first he needs to catch up to Waymo
  19. The latest AI models are hungry for inference compute
  20. Governor Newsom vetoed California's most controversial AI safety bill
  21. OpenAI just unleashed an alien of extraordinary ability
  22. 15 self-driving companies to watch
  23. Tesla's near-infinite training data isn't a silver bullet
  24. How transformer-based networks are improving self-driving software
  25. Human drivers are to blame for most serious Waymo collisions
  26. We could be months away from driverless semi trucks on Texas freeways
  27. What the AI safety movement can learn from the self-driving industry
  28. Grok 2 and Llama 3.1 don't stand out from the pack
  29. Waymo is growing exponentially
  30. A Tesla Model Y put me in the self-driving penalty box
  31. Predictions of AI doom are too much like Hollywood movie plots
  32. California should rethink its broad and sloppily drafted deepfake bill
  33. A California bill could seriously limit open-weight AI models
  34. Measuring intelligence isn't easy
  35. AI mailbag: what's going on with Alexa?
  36. Thoughts on Leopold Aschenbrenner's “situational awareness”
  37. Anthropic and Google's new mid-sized models are on par with GPT-4o
  38. How consumer demand could limit AI-driven growth
  39. The hidden side of Apple Intelligence
  40. Please ask me a question
  41. Why Substack launched a support chatbot
  42. Anthropic decoded the vectors Claude uses to represent abstract concepts
  43. Johansson's beef with OpenAI points to unclear laws on voice cloning
  44. Waymo’s investments in San Francisco may be paying off
  45. On self driving, Waymo is playing chess while Tesla plays checkers
  46. The new ChatGPT has a lot more personality (but only a little more brains)
  47. Not-so-large language models: a concise guide
  48. The Ukraine war is driving rapid innovation in drone technology
  49. Grandes modelos de linguagem, explicados com o mínimo de matemática e jargão
  50. I'm taking a short break
  51. Amazon is dropping no-checkout tech from many grocery stores
  52. The fall of Inflection makes me more skeptical of "scaling laws"
  53. States are racing ahead of Congress to regulate deepfakes
  54. Live near DC or SF? Let's have a drink!
  55. Grandes modelos de lenguaje (LLM): una explicación con un mínimo de matemáticas y tecnicismos
  56. Claude 3: ChatGPT finally has a serious rival
  57. Why Elon Musk sued OpenAI
  58. Google's newest LLM can handle 10 million tokens of context
  59. Gemini Advanced is not that advanced
  60. Why The New York Times might win its copyright lawsuit against OpenAI
  61. The coming Waymo monopoly
  62. The case for—and against—rapid AI-driven growth
  63. A DeepMind AI rivals the world's smartest high schoolers at geometry
  64. FunSearch, self-driving trucks, and how to keep up with AI news
  65. AI could be bad news for young lawyers
  66. What's ahead for large language models?
  67. Why I'm skeptical simulation can enable superintelligence

2023 32

  1. Why the New York Times sued Microsoft and OpenAI
  2. New data shows Waymo crashes a lot less than human drivers
  3. Ask me anything
  4. Google Gemini and the future of large language models
  5. The real research behind the wild rumors about OpenAI’s Q* project
  6. What the heck happened at OpenAI?
  7. A lot of jobs aren't in danger from AI
  8. I just activated paid subscriptions and would love your support
  9. Firing Sam Altman hasn't worked out for OpenAI's board
  10. Kyle Vogt exits Cruise, leaving Waymo as the clear robotaxi leader
  11. Why I'm not afraid of superintelligent AI taking over the world
  12. OpenAI's vision: a chatbot in every app
  13. Joe Biden's ambitious plan to regulate AI, explained
  14. California suspension is an existential threat to Cruise
  15. Ubiquitous cameras plus AI will change how we understand the world
  16. We need to talk about facial recognition
  17. Don’t exclude AI-generated art from copyright
  18. Regulating AI won’t be easy
  19. Do driverless cars have a first responder problem?
  20. Driverless cars may already be safer than human drivers
  21. Large language models, explained with a minimum of math and jargon
  22. How human translators are coping with competition from powerful AI
  23. AI Mailbag: the risks and benefits of AI
  24. Ask me anything!
  25. The death of self-driving cars is greatly exaggerated
  26. Congress shouldn't rush into regulating AI
  27. The AI safety debate is focusing on the wrong threats
  28. Hollywood writers strike to keep AI out of their industry
  29. Why I'm not worried about AI causing mass unemployment
  30. I cloned my voice with AI and my mother couldn’t tell the difference
  31. Copyright lawsuits pose a serious threat to generative AI
  32. Let's learn about AI together