Dan Hendrycks Profile picture
Oct 16 5 tweets 3 min read Read on X
The term “AGI” is currently a vague, moving goalpost.

To ground the discussion, we propose a comprehensive, testable definition of AGI.
Using it, we can quantify progress:
GPT-4 (2023) was 27% of the way to AGI. GPT-5 (2025) is 58%.

Here’s how we define and measure it: 🧵 Image
Image
Our definition of AGI is an AI that can match or exceed the cognitive versatility and proficiency of a well-educated adult.

To measure this, we assess the multiple dimensions of intelligence derived from the most empirically validated model of human intelligence (CHC theory). Image
Image
For example, testing shows that models get 0% on Long-Term Memory Storage (continual learning).

Without persistent memory, current AIs have “amnesia.”

Relying on massive context windows is a “capability contortion”—a workaround that masks this fundamental limitation. Image
People who are bullish about AGI timelines rightly point to rapid advancements like math.

The skeptics are correct to point out that AIs have many basic cognitive flaws: hallucinations, limited inductive reasoning, limited world models, no continual learning.
There are many barriers to AGI but they each seem tractable.
It seems like AGI won't arrive in a year, but it could easily this decade.

Website: agidefinition.ai
Paper: agidefinition.ai/paper

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Dan Hendrycks

Dan Hendrycks Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @DanHendrycks

Mar 27
For the record I do not bet on this multiyear research fad.

To my understanding, the main way to manipulate the inner workings of AI is representation control. It's been useful for jailbreaking robustness, finetuning resistant unlearning, utility control, model honesty, etc. 🧵
Here's @andyzou_jiaming discussing limitations of bottom-up approaches to transparency
Papers that have made measurable progress with representation control:
circuit-breaker.ai (circuit breakers)
wmdp.ai ("unlearning")
tamper-resistant-safeguards.com (tamper resistance)
emergent-values.ai (utility control)
mask-benchmark.ai (LoRRA for lying)
Read 4 tweets
Mar 5
Superintelligence is destabilizing.

If China were on the cusp of building it first, Russia or the US would not sit idly by—they'd potentially threaten cyberattacks to deter its creation.

@ericschmidt @alexandr_wang and I propose a new strategy for superintelligence. 🧵 Image
Image
Some have called for a U.S. AI Manhattan Project to build superintelligence, but this would cause severe escalation. States like China would notice—and strongly deter—any destabilizing AI project that threatens their survival, just as how a nuclear program can provoke sabotage. Image
This deterrence regime has similarities to nuclear mutual assured destruction (MAD). We call a regime where states are deterred from destabilizing AI projects Mutual Assured AI Malfunction (MAIM), which could provide strategic stability. Image
Read 8 tweets
Feb 11
We’ve found as AIs get smarter, they develop their own coherent value systems.

For example they value lives in Pakistan > India > China > US

These are not just random biases, but internally consistent values that shape their behavior, with many implications for AI alignment. 🧵 Image
Image
Image
As models get more capable, the "expected utility" property emerges---they don't just respond randomly, but instead make choices by consistently weighing different outcomes and their probabilities.
When comparing risky choices, their preferences are remarkably stable. Image
We also find that AIs increasingly maximize their utilities, suggesting that in current AI systems, expected utility maximization emerges by default. This means that AIs not only have values, but are starting to act on them. Image
Read 8 tweets
Jan 23
We’re releasing Humanity’s Last Exam, a dataset with 3,000 questions developed with hundreds of subject matter experts to capture the human frontier of knowledge and reasoning.

State-of-the-art AIs get <10% accuracy and are highly overconfident.
@ai_risk @scaleaiImage
Image
Image
Image
Paper, dataset, and code: lastexam.ai
NYT article: nytimes.com/2025/01/23/tec…
Spot errors with the dataset? Correction form here: docs.google.com/forms/d/1M6djW…
Read 4 tweets
Sep 9, 2024
We've created a demo of an AI that can predict the future at a superhuman level (on par with groups of human forecasters working together).
Consequently I think AI forecasters will soon automate most prediction markets.

demo:
blog: forecast.safe.ai
safe.ai/blog/forecasti…



Image
Image
Image
Mock-up. Forecasting is _not_ currently integrated into X.
The bot could have been called "Nate Gold," but I didn't get permission from @NateSilver538 in time; hence it's FiveThirtyNine instead
@PTetlock @tylercowen @ezraklein @kevinroose
Read 5 tweets
Jun 6, 2024
California Bill SB 1047 has been amended after significant public discussion and negotiations between many different stakeholders. They preserve most of the benefits of the bill and fix some potential loopholes. 🧵
#1: The bill updates the threshold for a model to be covered. Some have argued that since compute gets cheaper over time, the 10^26 threshold would soon cover small startups. This criticism was overblown before, and even more so now, with the additional $100M threshold. Image
With the amendments, if a model costs less than $100M to train, developers have no extra obligations. This could mean hazardous models end up not being covered over time, but I expect that hazardous capabilities will arise first in $100M+ models; precision traded off with recall.
Read 9 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(