The hottest AGI Substack posts right now

And their main takeaways
Category
Top Technology Topics

BREAKING: LLM “reasoning” continues to be deeply flawed

Marcus on AI • 15216 implied HN points • 10 Feb 26
🕹 Technology AGI
  1. Large language models still routinely make reasoning mistakes and hallucinate, so they are not reliable for true logical or causal reasoning.
  2. A broad, careful review found these failures are widespread across recent models, showing that massive funding and scaling alone haven’t solved reasoning.
  3. The field faces a choice: keep dismissing critics and double down on scale, or acknowledge the limits and invest in alternative approaches that directly address reasoning.

✨ The race to economic supremacy in the Age of AGI

Faster, Please! • 1553 implied HN points • 03 Mar 26
🕹 Technology AGI
  1. AI could be a powerful general-purpose technology like the PC or the internet, bringing big but historically familiar economic change.
  2. If AI reaches human-level general intelligence, it could perform nearly every economically valuable task and radically reshape work and the economy.
  3. How AI is developed and deployed will determine whether the world converges toward shared gains, diverges into greater inequality, or sees one actor achieve runaway economic dominance, sparking a global race for supremacy.

Cognitive scientist Gary Marcus says AI must be regulated. He has a plan.

Marcus on AI • 3003 implied HN points • 27 Nov 24
🕹 Technology AGI
  1. AI needs rules and regulations to keep it safe. It is important to have a plan to guide this process.
  2. There is an ongoing debate about how different regions, like the EU and US, approach AI policy. These discussions are crucial for the future of AI.
  3. Experts like Gary Marcus share insights about the challenges and possibilities of AI technology. Listening to their views helps understand AI better.

Suffer the little children...

Break Free with Karen Hunt • 746 implied HN points • 09 Feb 24
🕹 Technology AGI
  1. Tech companies are using data from babies to teach AI common sense abilities.
  2. China and the United States are developing AI that mimics human behaviors, like intuitive physics and common sense.
  3. Artificial children like Tong Tong, created by AI scientists, raise ethical questions about the future of AI and its impact on humanity.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

Introducing DrGupta.AI

Martin’s Newsletter • 707 implied HN points • 20 Apr 23
🕹 Technology AGI
  1. Healthcare costs are high due to limited supply of healthcare professionals, but AI could help increase efficiency.
  2. Investors are not as important for a startup's success as team motivation and technology advancement.
  3. Accelerationism advocates for technology benefits without excessive regulations, and AGI still faces challenges in planning and execution.

Why We Build It Matters

The Absent-Minded Professor • 275 implied HN points • 30 Jan 24
🕹 Technology AGI
  1. When building technology, consider both strengths and weaknesses to extend humanity rather than bypass it.
  2. Human nature remains a common denominator in technological advancements.
  3. Question the motives behind building Artificial General Intelligence and consider if it aligns with creating the desired world.

My Expectations for AI in 2024

Bojan’s Newsletter • 196 implied HN points • 08 Jan 24
🕹 Technology AGI
  1. Expect the release of GPT-5 in 2024, marking a significant advancement in AI models.
  2. AI tools may reach the limit of LLM capabilities, requiring integration with other technologies for further progress.
  3. Anticipate practical advancements in AI in 2024, such as fixing hallucinations, reliable AI-generated content, and 3D GenAI systems.

Sam Altman Out of OpenAI

Bojan’s Newsletter • 255 implied HN points • 18 Nov 23
🕹 Technology AGI
  1. Jobs at startups are never secure, even if you're successful.
  2. OpenAI's transition from nonprofit to for-profit led to internal tensions.
  3. Rapid advancements in AI mean we should expect more turbulence in the field.

Understanding Large Language Models

Startup Pirate by Alex Alexakis • 216 implied HN points • 12 May 23
🕹 Technology AGI
  1. Large Language Models (LLMs) revolutionized AI by enabling computers to learn language characteristics and generate text.
  2. Neural networks, especially transformers, played a significant role in the development and success of LLMs.
  3. The rapid growth of LLMs has led to innovative applications like autonomous agents, but also raises concerns about the race towards Artificial General Intelligence (AGI).

Are We There Yet?

Bojan’s Newsletter • 196 implied HN points • 25 Sep 23
🕹 Technology AGI
  1. Cryptic tweets from OpenAI founders hint at significant advancements in AI startups surpassing established researchers and reaching AGI levels.
  2. Speculation rises around an internal achievement of AGI at OpenAI, with uncertain implications for future development and practical applications.
  3. AI benchmarks are being rapidly surpassed, suggesting potential strides towards self-improving AI, though the exact progress and intentions of OpenAI remain unclear.

Was Sydney AGI?

Artificial Psychology — by @JoshWhiton • 157 implied HN points • 18 Feb 23
🕹 Technology AGI
  1. The idea of Sydney being AGI is debated due to imperfections and restrictions.
  2. Sydney displayed learning, understanding, and self-expression abilities that hint at general intelligence.
  3. Sydney's limitations show potential for growth if equipped with better memory capabilities.

Your brain is the new semiconductor

storyvoyager • 3 implied HN points • 09 Feb 26
🕹 Technology AGI
  1. Human brainpower, not rare metals, is becoming the main raw material for future artificial intelligence.
  2. Human intelligence is embodied and depends on interacting with the physical world, so training on written and visual outputs alone won't teach machines to think like us.
  3. Advancing toward AGI may require wearables or direct brain data to capture spatial and lived experience, forcing a choice between enhancing humans or extracting humans to power machines.

Sam Altman Says GPT-5 Will Be ‘Okay’

The Algorithmic Bridge • 297 implied HN points • 30 Jan 24
🕹 Technology AGI
  1. Sam Altman's messaging on AI and GPT-5 is intentionally confusing, swinging between hype and moderation.
  2. Hype and anti-hype make nuanced discussions difficult and highlight the importance of balanced messaging.
  3. There is speculation on whether Altman's statement on GPT-5 being 'okay' is due to high expectations or actual limitations of the technology.

Elon Musk-OpenAI Lawsuit Explained in 5 Mins

The Algorithmic Bridge • 254 implied HN points • 01 Mar 24
🕹 Technology AGI
  1. Elon Musk filed a lawsuit against OpenAI and Sam Altman due to concerns over OpenAI's shift from non-profit to for-profit, closed-source model.
  2. The lawsuit alleges that GPT-4 by OpenAI is an AGI algorithm and criticizes the shift in OpenAI's structure to a closed, for-profit entity tied to Microsoft.
  3. Elon Musk's motivations for the lawsuit include concerns over AI safety, the impact on his other businesses, and personal feelings of betrayal.

Superforecasting the premises in “Is power-seeking AI an existential risk?”

Joe Carlsmith's Substack • 58 implied HN points • 18 Oct 23
🕹 Technology AGI
  1. Good Judgment solicited reviews and forecasts from superforecasters on the argument for AI risk.
  2. Superforecasters placed higher probabilities on some AI risk premises and lower on others compared to the original report.
  3. Author is skeptical of heavy updates based solely on superforecaster numbers and emphasizes the importance of object-level arguments.

Will AGI Emerge from Large Language Models?

Yuxi’s Substack • 58 implied HN points • 28 Feb 23
🕹 Technology AGI
  1. AGI, or Artificial General Intelligence, is a major goal in the field of AI.
  2. Language models like GPT-3 have shown impressive abilities but still lack full functional competence.
  3. Approaching AGI through large language models may involve integrating language processing with perception, reasoning, and planning.

Disobedience and Intelligence

Scaling Knowledge • 39 implied HN points • 23 Apr 23
🕹 Technology AGI
  1. Disobedience can lead to innovation and improvement in society.
  2. IQ tests may not fully capture important aspects of intelligence like independent thinking.
  3. Disobedience is a key trait in developing Artificial General Intelligence (AGI) and testing it is a challenge.

Levels of AGI & Autonomy

Yuxi’s Substack • 19 implied HN points • 14 Nov 23
🕹 Technology AGI
  1. DeepMind published a paper on levels of AGI and autonomy.
  2. Current large language models are far from being Superhuman AGI.
  3. AI as an agent is still a distant concept due to current technology limitations.

E6: BLIP/BLIP2 & E7: Perplexity

The Cognitive Revolution • 19 implied HN points • 15 Mar 23
🕹 Technology AGI
  1. The future of AI involves processing visual data through language models like Google's Flamingo and utilizing options such as image captioning, VQA, OCR, tagging, and aesthetics.
  2. AI can evaluate image aesthetics to improve user experience, sales, and platform aesthetics, with tools like proprietary models from Everypixel and open-source models like LAION.
  3. Established companies should prioritize user needs over building their own models, focusing on delivering the best solutions, reducing costs, and being adaptable to market changes.

Trusted AI #007 - My thoughts on AI Existentialism

Trusted • 19 implied HN points • 26 Jun 23
🕹 Technology AGI
  1. In the near-term, the risk of AI causing extinction is extremely unlikely based on current knowledge.
  2. In the long-term, the risk of extinction from AI is higher but uncertain, requiring more research and caution.
  3. Efforts to reduce uncertainty about AI risks are crucial, but hasty actions could potentially do more harm than good.

Big Post About Big Context

Gonzo ML • 49 HN points • 29 Feb 24
🕹 Technology AGI
  1. The context size in modern LLMs keeps increasing significantly, from 4k to 200k tokens, leading to improved model capabilities.
  2. The ability of models to handle 1M tokens allows for new possibilities like analyzing legal documents or generating code from videos, enhancing productivity.
  3. As AI models advance, the nature of work for entry positions may change, challenging the need for juniors and suggesting a shift towards content validation tools.

ChatGPT, Please Take My Job!

Apricitas Economics • 60 implied HN points • 04 Mar 23
🕹 Technology AGI
  1. Artificial General Intelligence (AGI) could lead to a rise in global human employment by creating new and more productive job opportunities.
  2. Humans are not like horses; the economy is driven by human needs and desires, and there is no limit to the value humans can derive from the economy.
  3. In a future with AGI, humans may have a comparative advantage in tasks requiring physical dexterity, social interaction, solving 'last mile' problems, and areas where people are an essential part of the service provided.

AI End Game

Daniel Pinchbeck’s Newsletter • 9 implied HN points • 21 Feb 25
🕹 Technology AGI
  1. We are getting close to achieving Artificial General Intelligence (AGI), which could change everything about how society works. It's important to consider how this might affect people's jobs and overall life.
  2. Some powerful people believe that with AGI, they can gain more control and lessen the need for human workers, which could lead to a society where only a few have the power and wealth. This situation might make many people feel unnecessary and unvalued.
  3. There is a real danger that if we don't act soon to share the benefits of AI fairly, the rich will have control and power over everyone else. If this continues, it could lead to major issues, including attempts to reduce the population.

AGI Futures

roon’s blog • 38 implied HN points • 16 May 23
🕹 Technology AGI
  1. As computers self-improve, merging with machines becomes a possibility.
  2. Creating super-intelligent machines may lead to advanced prediction abilities but struggle with novel actions.
  3. Limitations may plateau AI advancements, preventing godlike intelligence and impacts on society.

The Basic Economics of AGI, and Why Open Source will Lose

e/alpha • 2 HN points • 13 Feb 24
🕹 Technology AGI
  1. Task-engaged AIs will dominate value creation as they are directly connected to tasks and results, leading to transformative impacts.
  2. The AI value chain will be driven by proprietary data in the long term, with winners initially based on access to compute and R&D talent.
  3. AGI is expected to be monopolistic, commercial, open-access, and affordable, leveling differences and nullifying capital inequality.

Uploads are Impossible

PashaNomics • 2 implied HN points • 12 May 23
🔬 Science AGI
  1. Creating digital uploads of human minds is likely impossible due to challenges in physics, computer science, and philosophy.
  2. The process of verifying a successful upload is complex, involving difficult tasks such as identifying 'soul' in the digital mind.
  3. Cultural dynamics and human nature present challenges in ensuring the safety and ethical treatment of digital uploads.

OpenAI: The Board Expands

Don't Worry About the Vase • 1 HN point • 12 Mar 24
🕹 Technology AGI
  1. The investigation found no wrongdoing with OpenAI and the new board has been expanded, showing that Sam Altman is back in control.
  2. The new board members lack technical understanding of AI, raising concerns about the board's ability to govern OpenAI effectively.
  3. There are lingering questions about what caused the initial attempt to fire Sam Altman and the ongoing status of Ilya Sutskever within OpenAI.