As I write this in mid/late 2026, a Doomerism narrative around the outcomes of AI development has gripped the nation, and the majority of people are starting to fear that AI is on course for the dystopian, rather than utopian, outcomes. Bans on Data Centers are spreading, politicians are calling for halts on super intelligence, and even executives in the leading labs are demanding government enforcement to prevent smart models from proliferating.

It’s a complicated situation, and there are various buried motives at play no doubt (curious coincidence that a government body enforcing model limitations would have the side effect of killing open source, which is growing very rapidly, as well as codifying an enormous moat around the few large players and blocking anyone else from entering). But it centers around a concept called RSI. Recursive Self Improvement.

Current teams of human engineers are working hand-in-hand with LLM models to improve each iteration of model. Improvements come from training techniques, training data quality, and other architecture optimizations, discoveries, and concepts. At some point, it is believed, AI will become smart and capable enough that the human-llm partnership will ultimately be slower and less efficient that a pure llm-development cycle. At this point the machine will be more capable of improving each of the inputs: training techniques, training data, and architecture. The resulting system, by definition, will be even more capable of improving its successor, and so on.

Once this happens, a feedback loop kicks off that pushes intelligence to infinity, vastly eclipsing humans, and by such a gap we can never assert control over the resulting thing again.

In my humble opinion, this is likely bullshit. I’ll put this forth as my asymptote theory. So here’s the thing: The current LLM architecture is trained on entirely on human knowledge (or derivative synthetic content, that comes from human-knowledge built systems), which means the models we select during training are the ones that have the fullest and best internal representation of that human knowledge corpus.

Right off the bat, there is no reason to think that such a process could exceed human knowledge in any given discipline… with two exceptions.

First, it would be expected that a perfect LLM model would reach a level of expert knowledge equivalent to the very best expert in a given field; however, unlike humans who are typically restricted to a single field of maximum expertise, the system could reach that equivalent across ALL fields. There are very few frontier level mathematicians who are also frontier level biologists. This is a huge advantage, sure, but it does not represent a path to infinity. Instead, it represents a raising of the asymptote to the cumulative sum of acquired human knowledge and the connections that might exist between disciplines that are hitherto undiscovered.

The second exception is time. You can run thousands of agents and do the equivalent of thousands of years of human work without ever exceeding human knowledge, and still accomplish incredible things. This is the case for OpenAI’s recent Navier–Stokes equation solve. They ran 10,000 agents over a hundred billion tokens, which was the equivalent of something like ~5,000 human years of research. It had no mathematical skill above human mathematicians, it just had a lot of time. There was nothing super-human about this, and humans probably could have solved it in much less time, but they nonetheless are outperformed by a less-intelligent system that has a massive time advantage.

So AI has the advantage of time leverage and the advantage of cross connections between disciplines. This is an advantage to be sure, but it’s still represents a hard asymptote at some low multiple above the sum of present human knowledge. Recursive self improvement is therefore not a recursive takeoff to infinity, it is a recursive approach towards an asymptote, which will result in ever diminishing returns.

What would change this is a breakthrough in architecture, or the ability to gather new knowledge by running novel experiments and reviewing the data. The first of those remains in the real of possibility, but there’s frankly no obvious path for how such a system could hit the point of recursion without humans in the loop. Remember it all starts with training and training data. Those runs are setup by people, and are not connected to the LLMs. Separate servers, separate workflows, separate everything. And the second of those is far out of the realm of possibility. When we have research labs setup with robots run by LLM systems without oversight, then it’s time to be worried, but at present every single step of the way requires human approvals, meetings with board members, politics, liability conversations with lawyers, physically plugging and unplugging things… thousands of steps that MUST move at human-time speeds and presently MUST involve humans.

The realistic scenarios where a singularity of RSI takes off to infinity and the system is able to jump from the digital realm to the physical realm in a non-trivial way is not realistic. Many major critical infrastructure systems have hard offline backups built in (left overs from cold war prep) that are explicitly non-hackable. Human manipulation is a real thing, but not some magic bullet either. Human systems have plenty of security built in to protect against malign actors.

The doomerism narrative is the kind of vague fear that we see whenever our status quo is changing in a big way, and it certainly gains wind from the many scifi stories that imagine our downfall to the machines, but realistically speaking, given the present composition of the world and our infrastructure and datacenters and AI development processes, it’s just not going to happen.

LLM Intelligence with RSI is unlikely to explode to infinity and more likely to find ever-diminishing returns towards an asymptote somewhere slightly above human knowledge. Even if massive breakthroughs are visible from that vantage, there is no path that lets it bypass humans in any real way to take over the world or kill us. Not until some far and imagined future where people have forgotten all these worries and handed the keys to our major systems to AI models, but that’s just not in the zeitgeist and not anywhere in the cards for decades.

Why does this matter? Because AI is actually a lot more than LLMs, and I really worry that in our present society where nuance is a forgotten concept, this is getting buried. In the rush to declare chat bots are going to destroy the world, we risk abandoning some real AI tools that have already brought a number of benefits, including:

• AlphaFold solved protein folding, which evaded human researchers for years, and has already aided in malaria vaccine research and will undoubtedly lead to hundreds of other therapeutic interventions that will save lives. Built on Evoformer, which has similarities to LLM’s architecture.
• MIT’s Mirai has enabled prediction of breast cancer up to 5 years earlier, which vastly improves outcomes. This exceeds human abilities, and other AI models are showing similar results on other cancer types.
• Accessibility for deaf/hard-of-hearing is undergoing a massive expansion with STT AI systems. YouTube can auto-generate captions, as can Zoom/MSTeams, and it’s coming to everything as AI has pushed the cost of this process so low. Transformers + CNNs.
• TREWS is an AI-powered (classic ML in this case, not transformers/NN) sepsis detection system now in wide use in hospitals. A two-year, multi-site study of 600,000 patients across five hospitals associated TREWS with a ~20% reduction in sepsis-related mortality because it could identify sepsis hours before doctors could see it.
• GenCast for disaster preparedness / extreme weather event prediction outperforms the European Centre for Medium-Range Weather Forecasts’ (ECMWF) Ensemble Prediction System (ENS) — the previous gold standard for this — 97% of the time. That’s a pretty big improvement. This is built on diffusion + transformers + attention (so essentially the same as the image generators like MidJourney).

This list goes on and on, and the ceiling for these kinds of targeted AI tools is not currently in sight. This stuff has real world beneficial impacts and real life saving outcomes that we are already seeing (not some vague techno-optimism).

Don’t give in to the doomerism. It’s not real. Asymptotes exist in the current architecture and training techniques. Fundamental breakthroughs in approach are possible, but there’s no evidence for that presently. Jumping from digital to physical is not some trivial step easily accomplished by a super-intelligence, unless we hand over a huge number of keys that nobody’s even considering handing to anything at present.

This is a superfluous post, with no purpose other than to commemorate the 10-year anniversary of this blog.

A lot has changed in the last 10 years, and thankfully all for the better. I’ve changed careers, moved three times, gotten married, had several kids, and even scribbled a novel!

Here’s to the next 10 years!

I’ve mentioned before, on this blog and elsewhere, what incredible value Brandon Sanderson’s YouTube lectures are to new writers. I had no formal training in writing, what with my background in Physics as well as Computer Engineering. It was the Sanderson lectures that helped me take a hobby and craft it into something more serious. The problem, I found, is they are split across four different channels, some are divided into pieces, years are skipped, and few have notes or details.

Several years ago I undertook the side project to gather all these lectures into one place. Things got in the way, and the project stalled a bit, but I have finally managed to post everything.

I tried to go a bit further than just mirroring the videos. I applied color correction to many of the series, as well as correcting the aspect ratio and combining separated parts of single lectures together:

Several of the later videos fell victim to my audio tinkering as well, in what I hope improved the sound quality (in 2014 in particular, the sources were sometimes rather badly lacking in audio).  Furthermore, for the 2016 “canonical” series, I’ve even transcribed the whiteboard in an overlay to the video to help translate the otherwise impenetrable “Sanderscript.”

Finally, I’ve gone to some lengths to take notes of the lectures and post them in the video descriptions, providing links and references from the lecture, and giving viewers the ability to jump to the relevant parts of the lecture:

This last step is what really held me up, and as the gap between videos yawned wider, I did crack a bit. For the 2014 and the latter half of 2013 series, I’ve gone ahead and uploaded them without any detailed notes. I fully intend to get around to them all in due course.

Oh, and of course, I didn’t monetize anything (and preserved the original Amazon affiliate links when I mirrored any such links from other channels).

So, here they are:

All lectures, in chronological order: https://www.youtube.com/playlist?list=PLSTzdBlo5WBDYapYaDq4A0cgQDk2CWhRD

Or, if you want to go series by series, here they are broken out:

The 2-part lecture series at JordonCon 2010: https://www.youtube.com/playlist?list=PLSTzdBlo5WBAdrQLWh6jJdu9BbE9ALmRT

BYU’s 2012 Creative Writing Semester: https://www.youtube.com/playlist?list=PLSTzdBlo5WBCsSpT-dN56bwbo5ig03cgR

BYU’s 2013 Creative Writing Semester: https://www.youtube.com/playlist?list=PLSTzdBlo5WBCM96UXr7GoLg5HLzP1G7uw

BYU’s 2014 Creative Writing Semester: https://www.youtube.com/playlist?list=PLSTzdBlo5WBAZRTMWz-9_CQbLXnZ-t40n

BYU’s 2016 Creative Writing Semester: https://www.youtube.com/playlist?list=PLSTzdBlo5WBB6ggp85NFzl7KNQa0ZdK_i

As far as I am aware, there is no published series for 2015 nor 2017, but I will keep an eye out. I did not collect any one-off lectures or short lectures that were part of book tours.

I sincerely hope this library helps you as much as it has helped me. Now….. Back to writing!

One of my many neuroses is an OCD-like organizing of my computer files, separated by year. I happened to be browsing some of the old folders, in this case from 2003, and I stumbled across this short I had written for a high school english assignment. I found it interesting enough to post. #throwbackthursday.

* * *

I remember the thunder and the rain hammering the roof like an invading army. The creaks and groans of the old wood like nearby battle cries. The windows blocked out the chill air, but were unable to stymie the darkness, which flowed in thick as smoke. The swaying, flickering, light cast by an overhead chandelier sent our huddled shadows dancing across the walls. I remember watching them move, as if they were strangers circling us, preparing to close in. Unbroken by the fearful barking of a dog forgotten outside, or the screech of tires as somebody rushed home, the hypnotizing rhythm of the rain and thunder threatened to consume the four of us entirely. We were usually talkative when we were together, but this evening we all sat silently, as if we somehow anticipated the moment that was about to arrive. What I don’t remember is what we did earlier in the day, or how we arrived in that room on that dark night. Nor can I resolve precisely what came after. That night exists for me only as the exaggerated minutes that preceded her confession.

Emi was looking particularly glazed; I suppose I should have recognized the slight frown on the edges of her mouth and her vacant eyes, and known something was very wrong. Regardless, we would all soon find out. Somewhere outside, lost in the darkness, a tree must have fallen. We all jumped when we heard the crash. I looked from Karen to Sarah, and from Sarah to Emi. It was in this moment as Emi’s expression was thrown into harsh relief, the darkness momentarily suspended by a burst of brightest blue, that I saw written on her face what I had failed to see before. I remember Sarah asked her if everything was all right, but Emi didn’t answer. The rain crashed down harder than ever, and the street lights outside flashed angrily and went out. Perhaps it was because she had planned to tell us, or perhaps something in the darkness of this desolate evening resonated with the weight upon her heart. I can’t say for sure, but I remember her eyes locked to mine, and there was a look of watery supplication I had never seen before. And a darkness, even darker than the cold air that surrounded us. Remembering the emptiness in those eyes still makes me shiver. She had stiffened and started speaking. At first her voice seemed stuck in her throat, but it finally found its courage, and she said she had to tell us something. She hesitated, and I found myself wishing she would not continue, but it was only a moments pause before she did. I have forgotten her exact words, but I remember the images her words summoned as if they were burned into my mind’s eye.

She told us how her boyfriend had called her to his house on that stormy night the previous weekend. Intent upon surprising him, she had arrived early, only to find a surprise of her own: he was there with another girl, sending her out into the cold after a loving embrace, then gesturing for her to depart before Emi arrived. There was a pause in her story, but none of us responded. Her words didn’t seem to settle, rather they hovered uncomfortably in the air between us, the darkness all around them. Emi explained that she had confronted him, and in her temper, she had taken his gun from his study and turned it on him. Apparently they struggled, and as he tried to fight her off, he cut her arm with a kitchen knife. Emi rolled up her sleeve and showed us. A flash of lightning timed itself dramatically and illuminated a jagged swollen cut across her forearm. Karen put her hands over her mouth and made a whimpering sound. Sarah looked like she was crying. I didn’t know what to think. But the rain hammered on, so I just sat and listened. The details are lost… I do not know if I’ve blocked them, or if she never told us, but as the struggle escalated, she shot him. At this point she was trembling all over, and her voice trailed off. So there it was: my best friend had killed someone. With a seeming crescendo, thunder erupted directly overhead and as the lightening retreated, the darkness was set free. But suddenly the darkness did not feel oppressive, it felt like a void. We were no longer surrounded, instead we were isolated afloat in nothingness, our senses paralyzed and blank. Sarah sat perfectly still, dim and distant lights glistening on her wide eyes. Karen turned to me but was unable to speak.

Something else was different between us now. The emptiness was leaking into my body from the surrounding void. I felt scared and lost; I didn’t know what to do. Karen asked Emi if she had called the police. Fighting to maintain her composure, Emi merely shook her head but did not speak. The thoughtful silence returned, and we all sat still, now four strangers in the dark. I saw the glimmer of silent tears running down Emi’s face when she sniffled. She had killed someone. Was she now a different person? Was I? I remember a part of me wanted to comfort her, but another part held me back. I saw Karen, too, looking at her differently. Sarah was looking the other way. The void was growing larger, a separation between us four as we sat there in the darkness, the blood from this dark secret seeping through us.

I don’t think any of us spoke for the rest of the evening; there was nothing to say. I don’t know if I was different or if Emi was different. It was a strange thing, that the actions of a few moments of emotion could change everything for her and for us. A few days later I believe Sarah tipped the police, and Emi was arrested. It’s funny how finite the world is. One moment things seem static: boundaries seem clear and reality is well defined. But then in the time it takes for rain to descend, a moments darkness changes everything. Sometimes I am able to feel bad for her, or at least for the person I knew and cared about before that night. I never found out if she was the same person. If the person I thought I knew was a lie, or if the darkness was never a part of her, but merely exploited her in a moment of weakness. Is evil real, or are we all arbitrary victims of inscrutable circumstance and inexorable darkness?