The image presents a line graph depicting global maternal mortality rates from 1985 to 2023. The y-axis represents the number of maternal deaths per 100,000 live births, ranging from 0 to 500, while the x-axis spans the years from 1985 to 2023. The graph shows a downward trend in maternal mortality rates, starting at approximately 460 in 1985 and declining steadily over the years. A notable increase is observed during the first year of the COVID-19 pandemic, but by 2023, the rate has decreased to about 197. This marks a 57% reduction compared to the rates in 1985. Annotations highlight key points, such as the spike during the pandemic and the significant decrease achieved by 2023. The data sources for this graph include the WHO, UNICEF, UNFPA, World Bank, and UN DESA (2025). There is a note stating that the estimates aim to adjust for underreporting and misclassification. The image is credited with a Creative Commons attribution.

Global maternal mortality rates have fallen by almost 60% since 1985

Inventing Humans - Substack
open.substack.com/pub/clues1/p...

A line graph depicting GDP per capita from 1820 to 2022, with the vertical axis representing GDP in international dollars (2011 prices) and the horizontal axis showing the years. Multiple colored lines represent different regions. The chart highlights a stark divergence: while average incomes in every region have increased, the pace of this growth has varied enormously. Western Europe and the “Western Offshoots” (like the US and Australia) experienced early and sustained economic growth. Meanwhile, Sub-Saharan Africa and South Asia grew much more slowly. Source: Bolt and van Zanden's Maddison Project Database. Chart CC BY Our World in Data

Global inequality is the result of two centuries of uneven economic growth

Always good to remember.

Safetensors and rust is the way🫡

The Angst created by our peers will not evolve to be true, unless we make it a self fulfilling prophecy.

'Intelligent' Machines will aid our creativity.

Intelligence is not what makes us human. An Olympiad does not become a goat 🐐 or rabbit🐇 by achieving incredible physical feeds.

We live in a state of Machine Angst. The age old story of fearing the unknown vs. inviting the adventure.

Erst kommt das Fressen dann die Moral.
--
First comes food, then morals.

Bertold Brecht

Folds in Flowers 🌺

Finding Fractals everywhere 🌺

Let's be kinder day by day! Do not hurry to get angry!

@lfcr.bsky.social We never walk alone. O, are those hearts broken. We will still lift trophies in May!!

Oh, people still need to do physical work? Bless those LLMs 😅

The AI data center boom will create new jobs. Most won't be at Amazon, Microsoft, Google, or Meta. Data centers contributed 4.7 million jobs to the US economy between 2017 and 2023, a PwC study found. Most of them aren't direct tech jobs.

Data centers contributed 4.7 million jobs to the US economy between 2017 and 2023, a PwC study found. Most of them aren't direct tech jobs.

In this generation of "Let's re-create humans." Are we creating Pinocchio's or Frankensteins?

So how do posts here work, exactly?

Emergent Symbolic Mechanisms Support Abstract Reasoning in Large Language Models Many recent studies have found evidence for emergent reasoning capabilities in large language models, but debate persists concerning the robustness of these capabilities, and the extent to which they ...

LLMs have shown impressive performance in some reasoning tasks, but what internal mechanisms do they use to solve these tasks? In a new preprint, we find evidence that abstract reasoning in LLMs depends on an emergent form of symbol processing arxiv.org/abs/2502.20332 (1/N)

#neovim Development News

The 0.11 version will have more built-in `vim.fs` file system functions (handy for plugin authors): `abspath`, `relpath`, `rm`.

PRs:
- github.com/neovim/neovi...
- github.com/neovim/neovi...
- github.com/neovim/neovi...

I don't like him invading wet / dry places.

Emergent Abilities of Large Language Models Scaling up language models has been shown to predictably improve performance and sample efficiency on a wide range of downstream tasks. This paper instead discusses an unpredictable phenomenon that we...

Reading "Emergent Abilities of Large Language Models".

Gut tells me: One new way of programming is finding the best fitting discontinuous function for a given domain.

arxiv.org/abs/2206.07682

How do you develop data frame driven Web Apps?

You are the C++ of our age.