Peter Lipman
LightNews LightNews
peterlipman.bsky.social
Peter Lipman
@peterlipman.bsky.social
1.6K followers 100 following 2.8K posts
I've been a teacher, a co-operative worker, a lawyer and external affairs director at UK charity Sustrans, and am the former chair of Transition Network, the Centre for Sustainable Energy and Common Cause Foundation.
Posts Replies Media Videos
Reposted by Peter Lipman
paigeblank.bsky.social
Don't miss this new @vox.com project that uncovers the rapidly changing world of fish farming.

Stories from @mbolotnikova.bsky.social + team challenge everything you think you know about the billions of aquatic animals that shape our planet’s future:

www.vox.com/future-perfe...
Eating the Ocean
How underwater factory farms are reshaping our food system.
www.vox.com
November 25, 2025 at 5:33 PM
Reposted by Peter Lipman
kaiheron.bsky.social
"The “ceasefire” is a diplomatic sham – a cover for the continuing extermination, displacement and erasure of the Palestinian people in Gaza and a distraction for the international public and the media." www.aljazeera.com/amp/opinions...
No, there is no ceasefire in Gaza
Israel’s bombing of Gaza is not a ‘violation of the ceasefire'. It is a continuing genocide under diplomatic cover.
www.aljazeera.com
November 25, 2025 at 5:44 PM
Reposted by Peter Lipman
carolecadwalla.bsky.social
I’m a past winner of the Orwell Prize & won partly because I exposed Vote Leave’s unlawful activities.

Michael Gove was its co-convener & refused to answer a single q. He’s now a judge of the prize & our world is truly one that Orwell would recognise

bylinetimes.com/2025/11/25/m...
Michael Gove Made Orwell Prize Judge Despite Record of Attacking Journalists and Dodging Scrutiny
Critics say "Orwell would have enjoyed the irony" of the former Conservative minister's appointment
bylinetimes.com
November 25, 2025 at 5:53 PM
Reposted by Peter Lipman
ketanjoshi.co
The more I think about this, the more it really, really, REALLY bothers me

We've reached a point where a text-based reference fabrication engine - comfortably the most effective machine liar we've ever invented - is being pushed by an ultra-polluting tech company and accepted by climate scientists
ketanjoshi.co Ketan Joshi @ketanjoshi.co · 1d
Deeply absurd. This Google PDF published on a blog (arxiv, not peer reviewed) claims an LLM is "PhD level" but in most cases the MAJORITY of reference URLs were invalid or inaccessible.

A PhD sitting down and just fabricating >50% of sources = career ending

arxiv.org/abs/2511.11597
CLINB: A Climate Intelligence Benchmark for Foundational Models Michelle Chen Huebscher1 , Katharine Mach2 , Aleksandar Stanić1 , Markus Leippold1,3, Ben Gaiarin1 , Zeke Hausfather4 , Elisa Rawat , Erich Fischer5 , Massimiliano Ciaramita1 , Joeri Rogelj6 , Christian Buck1 , Lierni Sestorain Saralegui1 and Reto Knutti5 1Google DeepMind, 2University of Miami, 3University of Zurich, 4Stripe, 5ETH Zurich, 6 Imperial College London Evaluating how Large Language Models (LLMs) handle complex, specialized knowledge remains a critical challenge. We address this through the lens of climate change by introducing CLINB, a benchmark that assesses models on open-ended, grounded, multimodal question answering tasks with clear requirements for knowledge quality and evidential support. CLINB relies on a dataset of real users’ questions and evaluation rubrics curated by leading climate scientists. We implement and validate a model-based evaluation process and evaluate several frontier models. Our findings reveal a critical dichotomy. Frontier models demonstrate remarkable knowledge synthesis capabilities, often exhibiting PhD-level understanding and presentation quality. They outperform “hybrid" answers curated by domain experts assisted by weaker models. However, this performance is countered by failures in grounding. The quality of evidence varies, with substantial hallucination rates for references and images. We argue that bridging this gap between knowledge synthesis and verifiable attribution is essential for the deployment of AI in scientific workflows and that reliable, interpretable benchmarks like CLINB are needed to progress towards building trustworthy AI systems. Total Reference URLs Generated claude-opus-4-1 claude-sonnet-4 gpt-5 hybrid gemini-2.5-pro gemini-2.5-flash o3 0.0 0.2 0.4 0.6 0.8 1.0 Proportion Reference URL Status hybrid gemini-2.5-pro claude-opus-4-1 o3 gemini-2.5-flash claude-sonnet-4 0 200 400 600 800 1000 Count of URLs Total Image URLs Generated hybrid claude-opus-4-1 gemini-2.5-flash gemini-2.5-pro claude-sonnet-4 o3 0.0 0.2 0.4 0.6 0.8 1.0 Proportion Image URL Status Status OK INACCESSIBLE_CONTENT INVALID_URL ERROR Figure 3 | Number of reference (top), and image (bottom), URLs and their status. Ablations We perform several ablation studies with the autorater (Table 4). Notably, removing the question-specific rubrics from the prompt changes the results only in the bottom half, with the Hybrid answers overtaken by Gemini 2.5 Flash and Claude Sonnet 4. This suggests that the additional resolution provided by the rubrics applies primarily to the kind of responses used to develop the rubrics. Or, in other words, that rubrics are far from complete. Hence, it is important that rubrics adapt to new data as better models become availab A New Expert-Grounded Benchmark for Scientific AI We introduce CLINB, a benchmark for modelbased evaluation of frontier models on complex, multimodal scientific communication. Its core is a new dataset of real-world climate questions paired with data-driven, question-specific evaluation rubrics, curated and validated by leading climate scientists through a novel three-phase, human-in-the-loop process.2 PhD-Level Synthesis vs. Attribution Failures Frontier models demonstrate remarkable knowledge synthesis, often exhibiting a PhD-level understanding. However, this performance masks a critical inadequacy in grounding. We report substantial hallucination rates for references (10% to 25%) and even more failures for images (50% to 80% in certain settings), exposing a major gap between synthesis and verifiable attribution. Insights into Human-AI Collaboration Dynamics Autonomous frontier models surpass ’hybrid’ answers (curated by experts using weaker AI assistance), revealing the assisting model’s capability—not human oversight—as the primary bottleneck. Counter-intuitively, highly motivated non-specialists (our ’Advocates’) who deeply engage with AI tools can produce higher-quality answers than domain experts who engage less with AI during answer curation. A Validated Methodology for Scalable Oversight We validate a rigorous, rubric-based autorater. Ablation studies demonstrate that structured prompts and automated evidence-checking are essential for mitigating inherent LLM judge biases. This process is hampered by inaccessible sources (up to 50%). Furthermore, we identify evaluation challenges, including model familiarity bias in human raters and the limitations of rubrics to generalize across models.
November 25, 2025 at 7:01 PM
Reposted by Peter Lipman
ketanjoshi.co
Deeply absurd. This Google PDF published on a blog (arxiv, not peer reviewed) claims an LLM is "PhD level" but in most cases the MAJORITY of reference URLs were invalid or inaccessible.

A PhD sitting down and just fabricating >50% of sources = career ending

arxiv.org/abs/2511.11597
CLINB: A Climate Intelligence Benchmark for Foundational Models Michelle Chen Huebscher1 , Katharine Mach2 , Aleksandar Stanić1 , Markus Leippold1,3, Ben Gaiarin1 , Zeke Hausfather4 , Elisa Rawat , Erich Fischer5 , Massimiliano Ciaramita1 , Joeri Rogelj6 , Christian Buck1 , Lierni Sestorain Saralegui1 and Reto Knutti5 1Google DeepMind, 2University of Miami, 3University of Zurich, 4Stripe, 5ETH Zurich, 6 Imperial College London Evaluating how Large Language Models (LLMs) handle complex, specialized knowledge remains a critical challenge. We address this through the lens of climate change by introducing CLINB, a benchmark that assesses models on open-ended, grounded, multimodal question answering tasks with clear requirements for knowledge quality and evidential support. CLINB relies on a dataset of real users’ questions and evaluation rubrics curated by leading climate scientists. We implement and validate a model-based evaluation process and evaluate several frontier models. Our findings reveal a critical dichotomy. Frontier models demonstrate remarkable knowledge synthesis capabilities, often exhibiting PhD-level understanding and presentation quality. They outperform “hybrid" answers curated by domain experts assisted by weaker models. However, this performance is countered by failures in grounding. The quality of evidence varies, with substantial hallucination rates for references and images. We argue that bridging this gap between knowledge synthesis and verifiable attribution is essential for the deployment of AI in scientific workflows and that reliable, interpretable benchmarks like CLINB are needed to progress towards building trustworthy AI systems. Total Reference URLs Generated claude-opus-4-1 claude-sonnet-4 gpt-5 hybrid gemini-2.5-pro gemini-2.5-flash o3 0.0 0.2 0.4 0.6 0.8 1.0 Proportion Reference URL Status hybrid gemini-2.5-pro claude-opus-4-1 o3 gemini-2.5-flash claude-sonnet-4 0 200 400 600 800 1000 Count of URLs Total Image URLs Generated hybrid claude-opus-4-1 gemini-2.5-flash gemini-2.5-pro claude-sonnet-4 o3 0.0 0.2 0.4 0.6 0.8 1.0 Proportion Image URL Status Status OK INACCESSIBLE_CONTENT INVALID_URL ERROR Figure 3 | Number of reference (top), and image (bottom), URLs and their status. Ablations We perform several ablation studies with the autorater (Table 4). Notably, removing the question-specific rubrics from the prompt changes the results only in the bottom half, with the Hybrid answers overtaken by Gemini 2.5 Flash and Claude Sonnet 4. This suggests that the additional resolution provided by the rubrics applies primarily to the kind of responses used to develop the rubrics. Or, in other words, that rubrics are far from complete. Hence, it is important that rubrics adapt to new data as better models become availab A New Expert-Grounded Benchmark for Scientific AI We introduce CLINB, a benchmark for modelbased evaluation of frontier models on complex, multimodal scientific communication. Its core is a new dataset of real-world climate questions paired with data-driven, question-specific evaluation rubrics, curated and validated by leading climate scientists through a novel three-phase, human-in-the-loop process.2 PhD-Level Synthesis vs. Attribution Failures Frontier models demonstrate remarkable knowledge synthesis, often exhibiting a PhD-level understanding. However, this performance masks a critical inadequacy in grounding. We report substantial hallucination rates for references (10% to 25%) and even more failures for images (50% to 80% in certain settings), exposing a major gap between synthesis and verifiable attribution. Insights into Human-AI Collaboration Dynamics Autonomous frontier models surpass ’hybrid’ answers (curated by experts using weaker AI assistance), revealing the assisting model’s capability—not human oversight—as the primary bottleneck. Counter-intuitively, highly motivated non-specialists (our ’Advocates’) who deeply engage with AI tools can produce higher-quality answers than domain experts who engage less with AI during answer curation. A Validated Methodology for Scalable Oversight We validate a rigorous, rubric-based autorater. Ablation studies demonstrate that structured prompts and automated evidence-checking are essential for mitigating inherent LLM judge biases. This process is hampered by inaccessible sources (up to 50%). Furthermore, we identify evaluation challenges, including model familiarity bias in human raters and the limitations of rubrics to generalize across models.
November 24, 2025 at 7:36 PM
Reposted by Peter Lipman
ketanjoshi.co
It has been a noticeable trend to see some renewable lobby groups becoming noticeably more right-wing, more pro-fossil fuels and more strongly in favour of deregulation and anti-environment laws if it helps get wind / solar built

Another recent EG I wrote about:

www.crikey.com.au/2025/10/20/q...
The Queensland Renewable Energy Council (QREC) sprinted in the opposite direction from muttered dissatisfaction, personally congratulating the state’s energy minister, posing for a pic, and poetically describing it as “a pragmatic and significant milestone balancing energy reliability with the need to unlock private capital to accelerate large-scale renewable investment”. It is a completely baffling statement, considering the fact that the plan assumes a hard zero new renewable energy development in the state beyond what’s already in the pipeline (the QREC is helmed by Katie-Anne Mulder, whose last job was 12 years at the Queensland Resources Council, the state’s main coal lobby group. I’m sure that’s irrelevant). The federal Clean Energy Council did not mention coal in its response, and didn’t offer an opinion on the plan as a whole.
November 25, 2025 at 7:25 PM
Reposted by Peter Lipman
ketanjoshi.co
"The growth of AI has been called the “savior” of the gas industry. In Virginia alone, the data center capital of the world, a new state report found that AI demand could add a new 1.5 gigawatt gas plant every two years for 15 consecutive years"

@ariellesamuel.bsky.social ---->
AI is guzzling gas
Big Tech is paying for gas plants and pipelines to directly power data centers, threatening global climate goals.
heated.world
November 25, 2025 at 7:39 PM
Reposted by Peter Lipman
leonsimons.bsky.social
I have very strong doubts about the 'Current Policy Scenario' (CPS) by the @iea.org.

What policies guarantee coal use to start a rapid decline within a decade?

Weak tea pledges don't count.

Unregulated fossil fuel supply has always found demand.
November 25, 2025 at 7:41 PM
Reposted by Peter Lipman
crisortunity.bsky.social
Amazing at this point they have even apologized for such turdulent journalism. Robbie Gibb must be furious
BBC Apologises for Failing to Challenge 'Dark Money' Climate-Denial Lobbyist
EXCLUSIVE: BBC faces fresh allegations of bias after corporation admits interview with ‘Net Zero Watch’ lobbying chief Andrew Montford 'fell below usual standards'
bylinetimes.com
November 25, 2025 at 6:08 PM
Reposted by Peter Lipman
paultherope.bsky.social
This is extremely serious. Why the last minute change? Are the government looking to replace Chamberlain with more gov friendly judges which the F35 case is another worrying example of? We need answers.
November 25, 2025 at 6:29 PM
peterlipman.bsky.social
Removal of judge in Palestine Action ban legal challenge ‘deeply concerning’ www.theguardian.com/uk-news/2025...
Removal of judge in Palestine Action ban legal challenge ‘deeply concerning’
Unusual last-minute change means panel of three will now hear case rather than Mr Justice Chamberlain
www.theguardian.com
November 25, 2025 at 5:20 PM
peterlipman.bsky.social
Google steers Americans looking for health care into "junk insurance" pluralistic.net/2025/11/25/o...
Pluralistic: Google steers Americans looking for health care into “junk insurance” (25 Nov 2025) – Pluralistic: Daily links from Cory Doctorow
pluralistic.net
November 25, 2025 at 5:13 PM
Reposted by Peter Lipman
leonsimons.bsky.social
Most don't realize that their are always billions of microscoping particles in the air around us.

Aerosols only become visible when the light and contrast are just right.

Without aerosols their wouldn't be cloud condensation.

Regional climate is impacted more by aerosols than by greenhouse gases
November 25, 2025 at 2:10 PM
Reposted by Peter Lipman
ninalakhani.bsky.social
Israel has killed two Palestinian children in Gaza every single day since the so-called ceasefire
November 25, 2025 at 3:05 PM
Reposted by Peter Lipman
davidsirota.com
🚨NEWS: Exxon is now petitioning the U.S. Supreme Court to grant corporations a new First Amendment right to avoid disclosing how much toxic pollution they are spewing and how much they are incinerating the atmosphere that supports all life on the planet.
Corporations Say It’s Their First Amendment Right To Hide
Corporations are asking the Supreme Court to help them dismantle a groundbreaking California transparency law requiring emissions disclosure.
www.levernews.com
November 25, 2025 at 4:24 AM
Reposted by Peter Lipman
opheliapg.bsky.social
Israeli Shelling Kills Civilian; Floods Ravage #Gaza Camps

imemc.org/article/isra...
Israeli Shelling Kills Civilian; Floods Ravage Gaza Camps
Israeli occupation forces killed a Palestinian Tuesday morning in shelling east of Khan Younis, in southern Gaza, while heavy rains at the same time flooded ...
imemc.org
November 25, 2025 at 3:07 PM
Reposted by Peter Lipman
thierryaaron.bsky.social
" #Climate change is hitting Thailand through water – and the time to act is now. The Hat Yai floods are not an isolated event but part of a worsening pattern of extremes"

#Thailand 🇹🇭

#ClimateCrisis
Another wake-up call: Hat Yai floods show Thailand is still unprepared for Climate Change
Recent severe flooding in Hat Yai highlights Thailand's ongoing struggle with climate change preparedness. Despite previous floods, extreme weather events continue to devastate the region.
www.nationthailand.com
November 25, 2025 at 3:07 PM
Reposted by Peter Lipman
ketanjoshi.co
Fantastic piece here by @Doctorvive.bsky.social

"Isn’t China becoming the world’s first electrostate, stepping into the role of global climate leader with the US embracing fossil authoritarianism? Well, China seems to be ambivalent too, at least for now"

www.theguardian.com/commentisfre...
Another Cop wrecked by fossil fuel interests and our leaders’ cowardice – but there is another way | Genevieve Guenther
The fingerprints of Russia and Saudi Arabia are all over the decision text in Brazil. But a group of nations led by Colombia and the Netherlands offer hope, says End Climate Silence founding director…
www.theguardian.com
November 25, 2025 at 4:15 PM
Reposted by Peter Lipman
cathobbs.bsky.social
Absolutely disgraceful 👇 - Labour's manifesto promised NHS would be "publicly owned and publicly funded"

Openly now ripping off patients for generations to come

Even the Conservative government said "goodbye to PFI"

Outrageous, unjustifiable, wrong
saulstaniforth.bsky.social Saul Staniforth @saulstaniforth.bsky.social · 22h
"Part of this will be capital funding we've allocated into the primary care estate.. & some of it, we're working with the private sector as part of a private sector partnership to fund the rest of them"

Minister Karin Smyth on how the new walk in health centres will be funded
November 25, 2025 at 4:22 PM
Reposted by Peter Lipman
weownit.org.uk
🚨BREAKING:

The government has announced they plan to use PFI 2.0 to fund NHS neighbourhood health centres. This will see cash that should go toward treating patients being paid out to private finance investors.

Email your MP now to oppose this plan:
weownit.org.uk/act-now/no-pfi
Just days to STOP new NHS private finance debts - take action now
Your MP can help STOP plans for new private finance debts in our NHS in the Autumn Budget
weownit.org.uk
November 25, 2025 at 10:10 AM
peterlipman.bsky.social
Most militaries report less than 10 percent of their carbon footprint www.sgr.org.uk/resources/mo...
Most militaries report less than 10 percent of their carbon footprint
Media release; 18 November 2025   A new report by Scientists for Global Responsibility (SGR) reveals that almost all official data on military greenhouse gas (GHG) emissions – even for those militarie...
www.sgr.org.uk
November 25, 2025 at 2:03 PM
Reposted by Peter Lipman
volts.wtf
A bunch of reactionary centrist pundits have decided that climate change isn't that big of a deal anymore. These climate scientists strongly disagree.

Hmmm ... how to adjudicate? 🤷
The world lost the climate gamble. Now it faces a dangerous new reality
The world bet on collective but voluntary action to keep global warming at a safe level.
theconversation.com
November 24, 2025 at 9:28 PM
Reposted by Peter Lipman
dalevince.com
Probably won’t see this in mainstream media - aka the right wing press - coz it’s against their narrative. Wind energy in Britain broke a new record last week with nearly 23GW on the grid one cold windy evening. That was 56% of all electricity on Britain’s grid.

buff.ly/r8yeFmJ
November 25, 2025 at 7:30 AM
Reposted by Peter Lipman
thierryaaron.bsky.social
"ETH researchers estimate the cost of removing 1 tonne of CO2 from the air in the year 2050 to be between 230 and 540 US dollars. This is twice as high as previous estimates"
thierryaaron.bsky.social Dr. Aaron Thierry @thierryaaron.bsky.social · 23h
"The cost of removing large quantities of CO2 from the air will fall in the medium term, but not as much as previously hoped.

This is the conclusion reached by @ethz.ch researchers on the basis of new calculations. Efforts to reduce carbon emissions should therefore continue at pace"
Cost of direct air carbon capture to remain higher than hoped
The cost of removing large quantities of CO2 from the air will fall in the medium term, but not as much as previously hoped. This is the conclusion reached by ETH researchers on the basis of new calcu...
ethz.ch
November 25, 2025 at 9:58 AM
Reposted by Peter Lipman
willnorman.co.uk
November 25, 2025 at 8:51 AM