Ramez Naam
Author and clean-energy analyst who writes on solar cost curves.
Ramez Naam did not write this page. ppll compiled it from their own words, each linked to the source. Is this you? · Report a mistake
Latest x.com
What they believe
-
Machines can be conscious
It's clear to me that consciousness can exist in other substrates than humans, including in digital systems.
x.comSaid 3 Oct 2026This on korrents
-
Changed their mind After four more doublings of installed solar capacity, solar electricity will cost about 3 cents per kwh in the sunniest areas and 4.5 cents per kwh in moderately sunny areas.
Before And beyond that, by the time solar scale has doubled 4 more times, to the equivalent of 16% of today’s electricity demand (and somewhat less of future demand), we should see solar at 3 cents per kwh in the sunniest areas, and 4.5 cents per kwh in moderately sunny areas.
Now Both of those forecasts were wrong. Solar has plunged in price faster than anyone – including me – predicted.
Said 14 May 2020This on korrents
The bold line is korrents’ wording; the quote is theirs. These one: what they wrote before what they said aloud, one per source, newest first.
Feed
As its own page →Hiding
16 October
15 October
14 October
13 October
11 October
10 October
3 October
From one piece @ramez on X 2 beliefs · x.com
-
korrents.com
It is far from clear that LLMs are conscious, because their architecture is very different from animal brainsTheir words
That doesn't tell us that LLMs are conscious, though. They have very different architecture than human or animal brains. It's far from a slam dunk.
-
korrents.com
Machines can be consciousTheir words
It's clear to me that consciousness can exist in other substrates than humans, including in digital systems.
27 September
From one piece Where’s the “intelligence explosion”? 3 beliefs · rameznaam.com
-
korrents.com
AI research at OpenAI is an order of magnitude or more harder than metrics, benchmarks, or forecasts suggest.Their words
A four-hour benchmark horizon is about 16 times longer than OpenAI’s research horizon. AI 2027’s 11-hour forecast is about 44 times longer. Of course, the tasks being performed by researchers at OpenAI aren’t the same as those in the METR benchmark. So we should expect some discrepancy. This, however, goes well beyond that. Actual AI research at OpenAI is an order of magnitude or more harder than metrics, benchmarks, or forecasts suggest.
-
korrents.com
Recursive self-improvement (RSI) may not lead to runaway superintelligence without a major conceptual breakthrough.AI alignmentrecursive self-improvement
Their words
I do expect autonomous self-improvement to arrive at some point. I’m skeptical that it leads to Type 5 - runaway super-intelligence - without a major conceptual breakthrough.
-
Their words
Given our best current data, the AI self-improvement loop would need to be roughly 5–10× stronger to sustain itself, let alone run away.
16 September
-
Their words
I can't rule out that there's some subjective experience happening in an LLM.
24 July
4 April
2 April
14 May 2020
-
Their words, before
And beyond that, by the time solar scale has doubled 4 more times, to the equivalent of 16% of today’s electricity demand (and somewhat less of future demand), we should see solar at 3 cents per kwh in the sunniest areas, and 4.5 cents per kwh in moderately sunny areas.
Their words now
Both of those forecasts were wrong. Solar has plunged in price faster than anyone – including me – predicted.
From one piece Solar's Future is Insanely Cheap (2020) 5 beliefs · rameznaam.com
-
korrents.com
Demand for electricity will rise to absorb far more solar than is currently considered feasible.Their words
Add these up, and it’s plausible that solar’s contribution to the energy system could double or triple the amount currently considered feasible for solar to provide.
-
korrents.com
Solar's future price should be predicted from how much has been built, not from how much time has passed.Their words
We can use the last decade to predict the future prices of solar power. Not by looking at the cost of solar as a function of time, but by looking at the cost of solar as a function of cumulative scale of the solar industry.
-
korrents.com
Average solar electricity prices in sunny parts of the world will fall to one or two cents per kwh by 2030 or 2035.Their words
This incredible pace of solar cost decline, with average prices in sunny parts of the world down to a penny or two by 2030 or 2035, is just remarkable.
+ 2 more
-
Their words
In short solar prices have dropped faster and lower than almost anyone has expected.
-
Their words
The data paint an incredible picture: One that shows the price of solar electricity from utility-scale systems dropping by anywhere from 30-40% with each doubling of cumulative solar deployment.
2 April 2019
19 September 2018
10 September 2018
5 January 2018
13 November 2017
15 August 2017
16 June 2017
25 March 2017
23 March 2017
21 November 2016
Nothing matches.
Beliefs
Korrents What they believe 11 beliefs, 1 changed mind — each backed by an exact quote.
Each is a — compiled by korrents.com, not by them: the one-line wordings are korrents', the quotes are theirs.
Recent
Machines can be conscious
It's clear to me that consciousness can exist in other substrates than humans, including in digital systems.
@ramez on X Said 3 Oct 2026
It is far from clear that LLMs are conscious, because their architecture is very different from animal brains
That doesn't tell us that LLMs are conscious, though. They have very different architecture than human or animal brains. It's far from a slam dunk.
@ramez on X Said 3 Oct 2026
Given current best data, the AI self-improvement loop would need to be roughly 5–10× stronger to sustain itself, let alone run away.
Given our best current data, the AI self-improvement loop would need to be roughly 5–10× stronger to sustain itself, let alone run away.
Where’s the “intelligence explosion”? Said 27 Sept 2026
Show 8 more
Recursive self-improvement (RSI) may not lead to runaway superintelligence without a major conceptual breakthrough.
I do expect autonomous self-improvement to arrive at some point. I’m skeptical that it leads to Type 5 - runaway super-intelligence - without a major conceptual breakthrough.
Where’s the “intelligence explosion”? Said 27 Sept 2026
AI research at OpenAI is an order of magnitude or more harder than metrics, benchmarks, or forecasts suggest.
A four-hour benchmark horizon is about 16 times longer than OpenAI’s research horizon. AI 2027’s 11-hour forecast is about 44 times longer. Of course, the tasks being performed by researchers at OpenAI aren’t the same as those in the METR benchmark. So we should expect some discrepancy. This, however, goes well beyond that. Actual AI research at OpenAI is an order of magnitude or more harder than metrics, benchmarks, or forecasts suggest.
Where’s the “intelligence explosion”? Said 27 Sept 2026
Some subjective experience in an LLM cannot be ruled out
I can't rule out that there's some subjective experience happening in an LLM.
@ramez on X Said 16 Sept 2026
Every doubling of the solar built worldwide has cut its price by 30 to 40 per cent, and that learning rate is why the price keeps falling.
The data paint an incredible picture: One that shows the price of solar electricity from utility-scale systems dropping by anywhere from 30-40% with each doubling of cumulative solar deployment.
Solar's Future is Insanely Cheap (2020) Said 14 May 2020
Solar has got cheaper faster than even its most optimistic forecasters expected.
In short solar prices have dropped faster and lower than almost anyone has expected.
Solar's Future is Insanely Cheap (2020) Said 14 May 2020
Average solar electricity prices in sunny parts of the world will fall to one or two cents per kwh by 2030 or 2035.
This incredible pace of solar cost decline, with average prices in sunny parts of the world down to a penny or two by 2030 or 2035, is just remarkable.
Solar's Future is Insanely Cheap (2020) Said 14 May 2020
Solar's future price should be predicted from how much has been built, not from how much time has passed.
We can use the last decade to predict the future prices of solar power. Not by looking at the cost of solar as a function of time, but by looking at the cost of solar as a function of cumulative scale of the solar industry.
Solar's Future is Insanely Cheap (2020) Said 14 May 2020
Demand for electricity will rise to absorb far more solar than is currently considered feasible.
Add these up, and it’s plausible that solar’s contribution to the energy system could double or triple the amount currently considered feasible for solar to provide.
Solar's Future is Insanely Cheap (2020) Said 14 May 2020
Beliefs others hold too
Machines can be conscious 2 hold this
It's clear to me that consciousness can exist in other substrates than humans, including in digital systems.
@ramez on X Said 3 Oct 2026
Every doubling of the solar built worldwide has cut its price by 30 to 40 per cent, and that learning rate is why the price keeps falling. 2 hold this
The data paint an incredible picture: One that shows the price of solar electricity from utility-scale systems dropping by anywhere from 30-40% with each doubling of cumulative solar deployment.
Solar's Future is Insanely Cheap (2020) Said 14 May 2020
Solar has got cheaper faster than even its most optimistic forecasters expected. 2 hold this
In short solar prices have dropped faster and lower than almost anyone has expected.
Solar's Future is Insanely Cheap (2020) Said 14 May 2020
What is a korrent?
A korrent is a belief a person has stated in their own words: one sentence stating the claim, backed by a quote and a source, kept at korrents.com.
Under a name here, the quoted block is what they actually said. The korrent beneath it is the claim those words support, in korrents' wording — tap it to see the record, its source, and who else holds it.
Nobody here wrote their own korrents. They are compiled from public statements, and a person can change their mind, which is recorded too.
Recommends rcmnd
rcmnd What they recommend 1 thing — each backed by an exact quote.
Where they publish
Newsletter rameznaam.com Their own newsletter on solar cost curves, energy and climate. Has a feed.
Long posts working through the economics of solar, storage and clean energy, with the numbers in the open — the series this site quotes for the solar cost-curve claims. Now published on Substack under their own domain, and it has a feed.
Recent
- Can AI self-improvement overcome diminishing returns? 27 Sept 2026 AI self-improvement may not be powerful enough to overcome diminishing returns. Even RSI may not lead to a fast takeoff to ASI (superintelligence).
- Two AI Futures to Choose From 24 Jul 2026 A messy, plural world of widely spread AI capabilities is a safer and more democratic one than that of a single, perfect AI.
- Two Hours of AI Optimism - Grounded in Reality 4 Apr 2026 My opening talk at Abundance 360, with Paige Bailey from Google. We covered the whole landscape of AI, how incredible it is, impacts on society, and some contrarian views on scaling, ASI, and AI hype.
Show 17 more
- Common AI Narratives are Wrong (Video and Part 1) 2 Apr 2026 In a recent talk I argue that many of the popular narratives on AI are wrong. Here's video and some of the key points.
- Solar’s Future is Insanely Cheap (2020) 14 May 2020 This is part 1 of a series where I’ll look at the future costs of clean energy and mobility technologies.
- The Third Phase of Clean Energy Will Be the Most Disruptive Yet 2 Apr 2019 Building new solar, wind, and storage is about to be cheaper than operating existing coal and gas power plants.
- Clean Technology is Disrupting Fossil Fuels Faster Than Ever 19 Sept 2018 The best news in the global fight against climate change is the pace at which clean technology is advancing.
- Yes, Blockchain Can Help Us Solve Climate Change – Why I Joined Nori 10 Sept 2018 Over the last two years, as both an angel investor and a public speaker on energy and climate, I’ve looked at least a few dozen startups and proposals on how to use blockchain or new crypto coins in energy or climate.
- Solar doesn’t need a “breakthrough”. It’s a breakthrough on it’s own. 5 Jan 2018 Yesterday, Tyler Cowan, who I’m a major fan of, wrote a piece for Bloomberg View arguing that solar needs more R&D for a true green energy breakthrough.
- Clean Energy Disruption – Video from South Africa 13 Nov 2017 I spoke at the SingularityU South Africa Summit earlier this year, about the way that solar, wind, storage, and electric vehicles are disrupting the $6 Trillion a year energy industry worldwide, and the opportunities fo…
- Don’t Let the Terrorists Win – White Supremacy Edition 15 Aug 2017 “Don’t let the terrorists win.”
- Why Trump Won’t be Impeached Any Time Soon 16 Jun 2017 I see any impeachment of Trump before 2019 as extremely unlikely.
- Healthcare Improvements Republicans Could Make 25 Mar 2017 Here are some things the GOP could productively do on healthcare, that have little or nothing to do with repealing the ACA:
- Our Future Cyborg Brains: My Keynote at XTech 2017 23 Mar 2017 I had a fantastic time keynoting the XTech Experiential Technology conference last week.
- To Fight Climate Change in the Trump Era, Focus on the States 21 Nov 2016 Summary: Focus on the states.
- Trump Isn’t Hitler. The US isn’t 1930s Germany 12 Nov 2016 Trump isn’t Hitler.
- We Can Push For Progress at the State Level 11 Nov 2016 Donald Trump won.
- New Record Low Solar Price in Abu Dhabi – Costs Plunging Faster Than Expected 22 Sept 2016 The price of solar power – in the very sunniest locations in particular – is plunging faster than I expected.
- Wind Power Blowing More Reliably Than Ever 17 Aug 2016 New wind turbines produce power more steadily – with less up and down intermittency – than ever before.
- How Cheap Can Electric Vehicles Get? 12 Apr 2016 This is part 5 of a series looking at the economic trends of new energy technologies.
Link verified 1 Sept 2026. Recent items update automatically from the channel.
x.com @ramez Posts daily on energy, climate and AI.
The short-form channel: solar and battery cost numbers as they land, AI, and climate argument. Higher volume than the newsletter and where most of the charts appear first.
Recent
- Looks pretty aligned to me. 16 Oct 2026
- When roon is positive about something in AI alignment or monitorability, I take that as a very good sign. 15 Oct 2026
- Reasons to be optimistic about AI alignment and controlability. 15 Oct 2026
Show 17 more
- Brendan's work is one of multiple reasons that I'm substantially less worried about AI hyper-persuasion now than I was last year. 15 Oct 2026
- Great articulation of something I've also been saying. AI capabilities are dependent on data: volume, accuracy, and ability to do perfect validation in compute (e.g., RL environments). First graphic is Alex's. 2nd is mine. 15 Oct 2026
- Is it good alignment or bad to consider the consequences of an action? Seems pretty good to me. Or at least, similar to one way we align human behavior. 14 Oct 2026
- Grounded, thoughtful takes on AI safety. 13 Oct 2026
- Progress Confernece was indeed one of the highlights of my year, both this year and last year. 13 Oct 2026
- As there's a chance I'm one of these people, I appreciate the call for grace. ;) 11 Oct 2026
- I feel compassion for those who's life work has been overtaken by AI. Perhaps I'll have this reaction one day. Let's not mock them. And: Mathematics isn't a jobs program for mathematicians. Advances in math are good, independent of the source. We need to find ways to adapt. 11 Oct 2026
- The Trump admin continues to weaponize every tool of government to go after its critics. We've learned the dangers of an overly powerful executive branch. Its control over the DOJ is one of many vectors of abuse. We need a constitutional amendment to make the DOJ independent. 11 Oct 2026
- In the AI safety debate I disagree with both maximalist camps. 1. Alignment-only camp: Everything comes down to alignment. You will never be able to build sandboxes, security sytems, or other safeguards that super-intelligence can't defeat. 2. Engineering-only camp: Alignment is a hoax. We need better and more secure systems, and nothing else. Why not work on both? Defense in depth has always been the way, and always will be. Build the onion of safety, many layers of independent and orthogonal ways to reduce the likelihood of and vulnerability to both accidents and malicious use. [To be clear… 11 Oct 2026
- We see this need for and positive impact of scaling very clearly in clean energy (and in compute, and now in AI). Invention is awesome. And scaling itself brings learning-by-doing, reducing the cost of new technologies and making it possible to disseminate them broadly. 11 Oct 2026
- This is an excellent and thoughtful post on managing AI systems. 11 Oct 2026
- Probably my favorite quote from my @MTSlive appearance yesterday. On the relative power of agent swarms vs more intelligent models. 10 Oct 2026
- Lots of fun chatting with @schisofrenia and @theojaffee about AI, RSI feedback loops, and why I don't see much evidence that we're close to runaway takeoff today. 10 Oct 2026
- I had a great time chatting with @JimPethokoukis about all things AI. - RSI and diminishing returns. - Narrow super intelligence and a continuing jagged frontier - AI and energy - Why I believe so strongly that broad democratization of AI is the safest path forward. 10 Oct 2026
- Completely agree. I'm shocked that Ed Boyden isn't included in this Nobel prize. 10 Oct 2026
- Excellent analysis from Toby. AI scaling with more tokens has historically been worse than how humans scale with more time to think. But that gap appears to be narrowing. Parameters like this are what we should be tracking to understand if we're headed for strong RSI 10 Oct 2026
- This gets to the heart of much discussion of ASI and AI risk. There are some domains where AI can become super intelligent solely with compute. Yet most important domains aren't like that. They face fundamental limits or real world data/ experimentation limits. Narrow ASI on highly verifiable domains now seems inevitable. Broad, general ASI does not. Maybe we'll get there. But that's far from certain. 10 Oct 2026
Link verified 1 Sept 2026. Recent items update automatically from the channel.
LinkedIn rameznaam Listed from their own site.
The professional channel, linked from their own site. Mostly the same energy and climate posts, aimed at an industry audience.
Link verified 1 Sept 2026.
About the English under a post
Some people here publish in a language other than English. Where they do, this site shows a machine translation beneath the post, in this typeface — the site's own, not theirs.
The post itself is never changed, moved or hidden: what is set in the serif above is exactly what the person published, and it is what to quote them on. A translation can be wrong in ways that matter, especially about tone.
Only the post's own words are translated. A quoted post, a linked article and a belief on korrents.com are left in their original language.