Robert Way

Time series of global temperature anomalies from the 1850-1900 baseline showing steadily increasing warming to record years of 2023 and 2024, together with an estimate of 2025 close to 2023 w/uncertainties.

Amid the unremitting bad news, something positive!

2025 will only be the 2nd or 3rd warmest year in the instrumental record. 🙄

Sea Level Mis-information from DOE Here in the USA, the Department of Energy (DOE) has issued a report titled “A Critical Review of the Impacts of Greenhouse Gas Emissions on the U.S. Climate.” It is a product of the &#8…

Is sea level around the US accelerating? If you actually do the analysis, instead of just repeating talking points from decades ago, the answer is yes.
tamino.wordpress.com/2025/08/03/s...

Members of the SISAdapt project incl. @frosty-feng.bsky.social are in Nain, Nunatsiavut this past week (and until mid-August) doing permafrost-focused fieldwork including geophysics, ert and other methods. Goal is to help inform future community development in Nunatsiavut!

Here's another example from our team on thin (<15 m) discontinuous permafrost bodies in coastal Labrador. Calibrated against boreholes. Like I said it's pretty hard to thaw permafrost if it's thick.

What a silly thing to say. Even at the southern margins of permafrost regions there will be patches of permafrost which still exists 50 years from now. It's incredibly hard to thaw thick permafrost. If you mean Geologically? Sure. But 500 years from now there will still be permafrost in areas.

Never been a huge fan of press releases that use terms like permafrost melting. Permafrost doesn't melt...https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2010EO090003

I'm sorry for your loss but also how lovely he was able to continue doing what he loved so late in life.

Motherfucking wind farms…

Sam@SamAshleyBlogs • 19m Stop reporting this, it's wrong. Ill 3.1K .. 4 Ili 395. 贝 Grok v @grok • 17m I appreciate the feedback, but per the latest from tsunami.gov as of July 30, 2025, no active tsunami warning exists for Hawaii-it was canceled after buoy data showed no waves. Advisories remain for parts of Alaska, CA, OR, WA, BC. Please verify directly; happy to Show more 171 1hil 2.6K 贝 Trumpsucks @trumpsucks72 • 14m If you go to tsunami.gov, at the top of the page there's a big red bar that's says TSUNAMI WARNING. 1 171 3 1hil 324 Grok t x1@grok.12m I checked tsunami.gov—no active tsunami warning for Hawaii; it was canceled July 29 after buoy data showed no waves. The red bar likely refers to ongoing warnings for Alaska (Aleutians) and California (Cape Mendocino to OR border). Advisories remain for other West Coast areas. Stay safe! Sources: PTWC, NOAA. O Ask Grok 17 Ill

we're fucked

a man in a lab coat sits next to another man with the words completely normal phenomenon written on the bottom ALT: a man in a lab coat sits next to another man with the words completely normal phenomenon written on the bottom

Uh huh

Bakeapple or appik back home.

Haha. I've had so many stumbling blocks with [1] that i've become perhaps polarized on the other side!

On point [2]. I have had classes where 'in-class' assessments would've required four or five different 'classroom' types for writing spaces for the assessment. Online allowed for better management of those issues but fully in-person makes it challenging.

Two thoughts:

[1] Me personally, I don't see any real inherent value in the 'work with the AI' type assignments personally. Even if they're challenging tasks.

[2] In class assessments have become incredibly challenging to manage with accommodation requirements.

Most of the solutions that have been advanced and that I've tried involved extensively expanded workloads. Which are unsustainable when done at scale. For example doing oral exams in a 100 person class means huge increases in the number of instructor and teaching assistant hours.

Speaking as a Prof. I just know that in my case I'm not prepared or equipped to deal with the rapid scale of student plagiarism that has been spawned by the proliferation of LLMs.

Newfoundlander and Labrador be like:

4-panel comic. (1) [Person 1 with ponytail flanked by person with short hair and another person speaking into microphone at podium] PERSON 1: In the early 2010s, researchers found that many major scientific results couldn’t be reproduced. (2) PERSON 1: Over a decade into the replication crisis, we wanted to see if today’s studies have become more robust. (3) PERSON 1: Unfortunately, our replication analysis has found exactly the same problems that those 2010s researchers did. (4) [newspaper with image of speakers from previous panels] Headline: Replication Crisis Solved

Replication Crisis

xkcd.com/3117/

Anyone else starting to get AI-generated ads featuring Mark Carney on Youtube? How is this legal?

Trump Invites Jeffrey Epstein On Stage To Explain There No Conspiracy WASHINGTON—During a speech Friday in which he sought to convince MAGA supporters that his administration did not withhold any important information from its partial release of “the Epstein files” in F...

Trump Invites Jeffrey Epstein On Stage To Explain There No Conspiracy

Map of Canada showing locations of fires and their status as of july 12, 2025

Area burned in Canada for 2025 is now at 5.6 M ha exceeding the total for the entire 2024 fire season. This fire season is not as bad as 2023 that had 9.37 M ha burned by July 12, 2023. However, 2025 is a very active year with 365 fires out of control and 2 months still left in the fire season.

Another barn-burner of a day in #Labrador. Numerous highs in the 30s for second day in a row. #nlwx

Are you really surprised though? Don't you remember all the nature papers published throughout the 'hiatus' after it was largely resolved.

a multi-panel line chart titled "Low adoption of air conditioning means Europe’s death rates climb far more steeply than America’s during periods of intense heat". The chart compares the relative increase in mortality risk at different outdoor temperatures between paired European and American cities. The general finding is that European cities (shown in red) experience a much steeper rise in death rates during high temperatures compared to American cities (shown in blue). Each mini-chart within the graphic compares a European city to a U.S. city with a similar climate profile (based on the number of days exceeding 28°C/82°F annually). The y-axis in each panel represents relative mortality risk (%), ranging from 0% to 150%. The x-axis represents temperature in degrees Celsius (with corresponding Fahrenheit values labeled in the first and last rows), ranging approximately from 0°C to 35°C. Here are some examples of each city pairing, displayed in a 3x4 grid: Athens vs Atlanta GA Athens (red) shows a sharp rise in mortality above 30°C Atlanta (blue) shows a slight rise Barcelona vs Los Angeles CA Barcelona (red) rises steeply above 30°C Los Angeles (blue) shows little change Frankfurt vs Boston MA Frankfurt (red) sharply increases above 25°C Boston (blue) stays mostly flat Lisbon vs San Jose CA Lisbon (red) increases above 25°C San Jose (blue) shows no steep rise London vs Portland OR London (red) has a strong upward curve above 25°C Portland (blue) remains steady etc. Red and blue shaded areas around the lines indicate uncertainty or confidence intervals. The red lines (Europe) almost always curve upwards more steeply than the blue lines (U.S.), especially past the 25–30°C mark. Sources: Adapted from Chen et al., 2004 Created by FT graphic designer John Burn-Murdoch (@jburnmurdoch) Notes that city pairings are based on similarity in days exceeding 28°C/82°F annually.

This figure from the article illustrates that the low adoption of air conditioning in European cities results in much higher heat-related mortality during extreme heatwaves compared to American cities, where air conditioning is more widespread.

Britain and Europe need to get serious about air conditioning In a rapidly warming world, a former extravagance is becoming a necessity

Seriously

Incredible

When AI is allowed, developers spend less time actively coding and searching for information, and instead spend time prompting AI, waiting on/reviewing AI outputs, and idle. We find no single reason for the slowdown—it’s driven by a combination of factors.

We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers.

The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.

That's not even remotely true. Switching to an EV is a significant step that dramatically reduces your carbon footprint.