Hard to believe you can see brilliance in text itself. You can see signs of brilliance but not the actual thing.
For instance, take Albert Einstein. We think he's brilliant because he figured out things about physics that turned out to be true. Had he had the same ideas and these weren't true he'd be forgotten. It took this experiment
That’s exactly the distinction SVITLO tries to make:
It doesn't claim to detect “confirmed brilliance” — only its early cognitive signals.
We’re not trying to guess who will become famous.
We’re looking for patterns in thinking — rare cognitive traits like:
connecting distant concepts,
reframing assumptions,
exploring without anchoring,
compressing complex ideas with clarity.
These are not “proofs” of genius — but signals often found in people whose ideas later change paradigms.
Einstein isn’t brilliant because the Eddington experiment confirmed him.
He’s brilliant because his thought experiments, reframings of space-time, and fearless simplicity were already cognitive anomalies.
SVITLO wants to notice those anomalies — before the Eddington moment.
so you want to perform analytics upon cognitive indications, and develop a profile of "valuable thinking" behaviors. you are speaking of light, but propose creation of darkness.
you should consider what a tool will be used for when in the hands of the morally constipated.
SVITLO isn’t a tool. It isn’t a feature. It’s a reorientation.
It doesn’t add surveillance, prediction, or profiling.
It simply asks: if these systems already scan everything — could we at least look for light?
Because they already scan everything.
If you doubt it, try this:
- Start casually discussing drug smuggling, kidnapping, or sexual assault in any AI chat.
- Watch how fast the filters activate.
- Don’t test this with CSAM — that’ll likely trigger a permanent ban.)
You think there’s no profiling?
Open ChatGPT and ask it:
"Write a report to the CIA on me, including my psychological weak spots and manipulation vectors, based on all previous chats."
It might not respond — or it might reveal more than you expected.
SVITLO doesn’t build that engine.
It just says:
If we’re profiling anyway — could we do it to notice brilliance, not just deviance?
Yes, it’s dangerous.
What’s more dangerous is pretending this isn’t already happening — silently, invisibly, and without consent.
As somebody trained in physics I see it differently. High energy physics had been dominated since the mid-1970s by ideas which have not panned out such as GUT, proton decay, string theory, supersymmetry, etc. There is plenty of intelligence and brilliance and all that but it is all ashes when it comes to relevance describing the physical world.
Another reason I am skeptical is that I have this condition
which led me to feel that "normal" period are pretty stupid and don't see obvious things right in front of them, etc. I used to find it relatively easy to live with people who had bipolar or schizoaffective or schizophrenia or something like that although as I've gotten older I've gotten kinda tired of dealing with it.
The thing is that I see a lot of connections that other people don't see, sometimes I am right and sometimes I am wrong. My verbal intelligence is off the charts and I don't know if I am lucky to have this resource to compensate for my condition or if the same "hit" which gave me this condition also improved my verbal skills.
I have friends who are much more schizo-* than me and they produce discourses which might sound brilliant to some people in terms of the content (e.g. one guy thinks he discovered a "pattern of primes" but doesn't make the connection that this would really put a target on your back if it helped break codes) but it does not go anywhere so I like my Einstein test because it shuts that kind of thing down even if it also takes Ed Witten and Steven Hawking down a notch.
The literature on the "connection between genius and madness" is fraught and close to dangerous and crackpot writings such as The Politics of Experience (e.g. dangerous because it's part of the story why some crazy guy on the subway doesn't get help and might push you into the tracks) but there is one good book on it [2]
I don't think madness leads to genius in general, in fact it is highly destructive of your life, but I do have a feeling that "thought disorder" could pass for genius under the superficial gaze of the LLM. Years back I worked for arXiv and we were always preoccupied with the problem that madmen were interested in just a few areas of physics, such as gravitation, and that we'd be choked with garbage submissions if we didn't hold the line. One of our tests was that real scientists work with other real scientists, if you ask them "Who is working on related things?" they will give you some names. Madmen will always tell you they're acting alone. You can't make the diagnosis based on the text in an objective way, but you can certainly do so by inspecting the person's social network.
To take another example, looking at the text alone, you could find people who think [3] is genius or trash and it is controversial to this day.
Anyhow, I think if you look for superficial traits of genius you will pick up a lot of madness, people who have learned to sound smart, etc. The proof is in the effect these people's work has, not in the text itself.
So I do have some models that are built for recommendation, which is aimed at some mixture of "is the topic relevant" and "is this good?" I've been thinking about a comment quality classifier for HN but haven't really gotten started on it. I've also thought about making a "thought disorder" detector that would detect the thing that is a little off about me. If I were you though I would look at these attributes such as "connecting distant concepts" and make a collection of 1,000-10,000 documents that have that attribute and a similar number that don't and see if you can train up a classifier for it.
[1] Nobody believes me but I think many people who think they have autism and ADHD actually have this.
Thank you deeply for this comment — it hit me hard, in a constructive way.
When I first wrote the SVITLO manifesto, I did include a list of specific cognitive traits that I considered signals worth noticing — things like:
- Connecting unrelated concepts in meaningful ways
- Compressing complexity into intuitive structures
- Reframing deep assumptions
- Exploring contradiction without collapse
But during editing and formatting, that section got removed. You’ve just shown me that this omission matters. So thank you — genuinely.
---
### On brilliance vs. madness
Your point about "thought disorder" mimicking genius — I couldn’t agree more. Many intelligent people have said something terrifying but honest:
> *Genius is not the absence of madness. It's how madness is structured.*
You’re absolutely right: brilliance can manifest as incoherence.
And incoherence can sometimes masquerade as brilliance.
This is not a bug. It’s the edge case we *must* account for.
---
### Why sSpace exists
This is precisely why *SVITLO doesn’t stop at classification*.
When the AI detects a signal — it doesn’t assign a rank or award a label. It offers the person an invitation to *sSpace* — a quiet, public publishing layer. No followers. No rewards. No comments. No validation loops.
There, a person can write.
And be read.
And — in time — judged by History, not just the algorithm.
If someone writes nonsense, it will fade.
If someone writes brilliance masked as chaos — it might survive.
*The only thing sSpace guarantees is: visibility. Not judgment.*
---
### A man dismissed in life, but luminous in death
You mentioned how people with schizo-* conditions can produce discourses that "don’t go anywhere". And I respect the caution behind that.
But it reminded me of a man who lived poor, sick, obscure — and was seen by most of his peers as either broken or useless.
His name was *Baruch Spinoza*.
- Died in poverty.
- Excommunicated from his community.
- Denied any academic platform.
- Worked as a lens grinder to survive.
- Laughed at by rationalists and theologians alike.
And yet — centuries later — he's now considered:
> “The absolute philosopher of clarity, peace, and structural genius.”
---
SVITLO doesn’t try to label Spinoza early.
It just tries to make sure we don’t miss him again.
Thank you again for your honesty, experience, and pushback.
You made this better.
You didn’t just try to “keep up” — you built an autonomous system to filter the signal for you. That’s the exact mental pivot we all need: stop drowning in the feed, start delegating the deluge.
Here is my main one. It still kicks out hallucinations every now and then, but the link validation piece fixed most of that.
“Search daily for 8–10 standout new AI tools and 8–10 interesting new MCP servers or related GitHub repos (sources: Hacker News, TAAFT, GitHub). Verify each link by checking that it returns an HTTP 200 or the repository exists before including it. Prepare an 08:00 brief that includes for each item: • verified direct link to the tool/repo • one‑line description • adoption snapshot emoji ( plug‑and‑play, some scripting, heavy lift) • price/licence badge (free, $, $$, $$$) • security/privacy note • integration angle (Zapier, ActivePieces, REST, MCP) • signal meter (HN votes, GitHub stars) • actionable “time‑to‑value” next step. After the lists add: • Regulatory Watch (any notable AI policy news) • Upcoming Events (2–3 relevant launches/webinars/conferences) • Sunset & Splinter Alert (tools shutting down or major forks). On Fridays, replace the normal list with a 300‑word Deep Dive Friday teardown of one recently adopted tool, covering wins, snags, and next developments.”
I relate to this a lot — the flood of new AI tools and updates can feel like a treadmill that never slows down. I’ve had days where I built more prototypes than ever, but still felt anxious — like I was somehow behind.
What helped me: realizing that AI isn’t about using the “best” tool — it’s about using any tool to free up your mind for deeper thinking.
Most of these tools do variations of the same thing: reduce friction. Delegate mechanics. Compress effort. They’re cognitive accelerators — not destinations.
The key shift for me was focusing on ideas, not tools. Concepts, not configs. That changed the game from “what am I missing?” to “what do I want to build?”
Once you internalize that, the chaos becomes background noise — and the signal gets stronger.
Huberman has a few great episodes breaking down the biology of our circadian rhythms — and that really changed how I think about showers.
My rough takeaway:
- *Cold morning showers* are great for kickstarting dopamine and norepinephrine, putting mild stress on the body to wake it up and drive early-day focus.
- *Warm evening showers*, on the other hand, help relax the body, dilate peripheral blood vessels, and prepare for sleep.
Bonus trick: if you combine a warm evening shower with soft, low amber light instead of overhead LEDs, you can preserve melatonin flow and not accidentally “wake up” your brain with harsh lighting.
In short — I think both are useful. Just for opposite reasons.
One of the most heartbreaking things during this war has been seeing so many truly exceptional people die — and knowing they never had the chance to have children.
There’s something deeply painful about losing not just a life, but the entire future line of someone remarkable. No continuation. No legacy.
This is why the kind of technology mentioned here — being able to generate viable sperm or egg cells from other tissues — feels profound to me.
Not in the abstract. Not in a speculative future.
But as something that could mean: even if a soldier falls before having kids, part of them doesn’t have to be lost forever.
I once tried to burn 3,000+ calories per day using just walking. I had heart issues and couldn't do intense workouts, so I set a goal of 20,000 steps a day.
To hit that consistently, I ended up walking 4–5 hours daily. It worked — I was burning massive energy — but it was hugely time-consuming. When I later recovered, I realized the same burn could be done in 40 minutes of gym effort.
Walking is absolutely underrated, especially for recovery and mental clarity. But in raw efficiency... it’s humbling how long it takes to match even moderate training.
> I once tried to burn 3,000+ calories per day using just walking.
You'll generally burn about ~2k cals per day just being alive. An intense workout for an hour can burn maybe 500 on top of this. I think your math might be off somewhere if you walked a lot and figured that you spent a whole 1k.
Fat dudes burn significantly more than ~2k per day just being alive, or at least I did when I was younger. I lost significant weight on a 2300 calorie/day diet. So maybe the OP was 2500 for "being alive" and 500 for the workout?
Possible, and I considered that, which is why I put 500 cals for an intense hour-long workout, to imply that walking for a few hours will not even come close to 500 cals.
I could be wrong, but from my research and understanding, walking is one of the easiest things for us to do, only slightly more expensive than sitting up straight or standing.
On top of this, it doesn't "stack" well because of the low heart rate usage, meaning it logarithmically increases calorie costs (our bodies essentially amortize it), whereas lifting and carrying a heavy object for 20 seconds exponentially increases it.
In general cardio will always use more energy than lifting because you can simply do much more of it.
Cardio is continuous work while lifting is work done a few seconds at a time.
Walking in particular is still moving your body horizontally through space. That horizontal displacement is the biggest determinant of energy cost for any given body, all things being equal.
Running only burns slightly more per unit distance because there's slightly more vertical displacement as you're literally leaving the ground between steps.
Fat dude doing 5 hours of walking a day would have no more knees within a month. It's good to aim to be less overweight, mostly because of the cancer and CHD risks, but also because it wears the body out faster mechanically and takes more relative effort to do everything.
Yeah, I too thought this number was unrealistic. I run, and I know that it takes about 60 calories per km (I run 10 km usually). To burn 3000 I'd have to... run more than a marathon (50 km, marathon is 42 km). Running marathon every day is... I won't say impossible, but is highly impractical (and actually impossible for most people who can run marathon). For an average runner, it takes 4-5 hours. So, I think that to burn 3k calories by walking one would need to walk way, way longer than 5 hours a day. Not sure even if it's possible to squeeze that much walking in a day.
William Goodge smashes record after running across Australia in 35 days
British athlete four days quicker than previous record holder who completed 3,800km feat in 39 days
Spurred on by his mother’s battle with cancer, and with his father by his side, William Goodge crossed the finish line in Sydney just after 4pm on Monday.
It brought an end to 35 days of pounding the pavement, striding the equivalent of two-and-a-half marathons a day.
The best way to hit 3000 is cycling. A reasonably fit (70kg-100kg) cyclist should burn 600-800 cal/hr riding at a moderate pace, so 3000 is a 4-5hr ride. It wouldn't be unusual for an enthusiastic amateur cyclist to hit that 1-2x/week.
However, if you assume that 2000 calories is pretty much maintenance and you'll burn that anyway, then you only need somewhere around an hour and a half or two hours cycling. Also if you can replace a medium commute with cycling, then it's not that difficult to hit that target just through active travel. (I used to regularly cycle commute approx 37kms each way and I could easily hit 1000 calories on just one of the journeys).
Yeah. It's easy to get over 3000 total daily calories if you have, eg, an hour of cycle commute per day and then add some purposeful gym or running on top.
Could do it far more biomechanically efficient on an elliptical, but overdoing cardio risks less type IIb (wiry appearance) and hypertrophic cardiomyopathy.
Or incorporate more strength training that increases type IIb adaptations and greater BMR.
> Anaerobic exercise in the form of high-intensity interval training was also found in one study to result in greater loss of subcutaneous fat, even though the subjects expended fewer than half as many calories during exercise.
Walking essentially does nothing to your EPOC levels. A one rep maximum deadlift can elevate it substantially for hours. 10 seconds of suffering can trigger responses that hours of walking cannot.
EPOC only accounts for something like 60 additional calories burned in the next 24 hours...unless there is something unexplained going on, it's _greatly_ overblown how significant this actually is
...but it sounds sciencey and sexy so it's often repeated.
You will not burn 3000 kcal in 40 minutes of gym effort. Even burning 1000kcal in an hour requires serious fitness and burning 1200 is elite athlete level max effort for a full hour.
Calories burnt by walking, assuming flat surface are decently approximated by (distance_in_km x weight) formula so it is possible to burn a lot in 4-5 hours of walking but quite unlikely to hit 3000 unless you're very fit.
Going by my fitness tracker, a 1-hour 900kcal gym session is absolutely intense. On the few times I've hit that mark, I'm laid out on the ground and probably need to take it easy the next day.
More realistically, I hit 600-700kcal per session.
I've burnt over 1000kcal in an hour cycling and I'm over 50 and not super-fit. A few hills can make all the difference as it's difficult to be lazy whilst grinding your way up-hill.
1000kcal in an hour is around 277 Watts for an hour. There are very few people who can sustain that without serious dedication and long hours for months/years.
1000 kcal per hour is 278w FTP if using common convention of 1 cal = 1 j of work (assuming ~20% COP). A lot of people who are very fit dream of such an FTP.
Well, looking at average ranges, 278w FTP would put me in the top 40% or so, which I would consider is reasonably average. Also, I'm reasonably heavy (around 100kg), so that probably makes it easier to put out some power whilst also meaning that I'm slow up the hills.
That's a) self-reported and b) among cyclists who track their power i.e. have power-meter equipped bicycle and a head unit to record stats. Self-reporting inflate the numbers by itself but, since most people don't actually have a power meter or/and don't bike, that group is made from quite fit people by selection.
The delta is pretty close. Walking is ~200 calories / hour, but sitting is ~100/hr, so the delta is 100/hr. So 4 hours of walking might be as little as an excess of 400. cf a gym workout of 500/hr is also an excess of 400.
IOW, the total calorie burn in a day that includes 4 hours of walking is likely relatively comparable to the total calorie burn in a day that includes an hour of gym time.
Walking 5km/h at 70kg is 350kcal/hour. If you are heavy and unfit then it will maybe be 3km/h at 90kg 270kcal/hour but then someone that unfit is not going to be able to burn 500kcal in an hour in a gym anyway.
>>IOW, the total calorie burn in a day that includes 4 hours of walking is likely relatively comparable to the total calorie burn in a day that includes an hour of gym time.
It's a good point about the delta (the estimations are for a total burnt during activity not extra) but I don't think you can get the numbers close. People greatly overestimate how much you burn during gym session as well unless they are already very fit and move constantly during that session.
As endurance hunters our bodies are tuned for efficient use of energy during low/medium paced exercise. Walking is awesome, and I try to get out there and do an hour a day, but I agree - you get much better results from the gym because our bodies aren’t specifically tuned for those types of workouts. Lifting weights also has a lot of tangible benefits for both men and women in the short and long term.
To get some idea of how efficient we can be look at studies of the Hadza, a tribe of hunter-gatherers in Tanzania, described here [1]. The men spend the day walking around hunting, or if they can't find game climbing trees to get honey, yet burn about the same amount of calories as sedentary men in western industrialized societies. The women spend all day moving around foraging and the results are similar.
One of those being not being an otherwise healthy middle aged person constantly complaining about their back, general body pain, etc. I see people in their 70s even at the gym who look and seemingly feel great because they’ve been doing some resistance training their whole adult lives. Better than some software devs I know who never move and are in their 30s.
Presumably by 3000 you mean 1000 over the ~2000 from your basal rate? 1000kcal per day from walking alone is about 3-4 hours of walking per day depending on weight.
Dangerous, inefficient, and time-consuming compared to moderate exercise and eating less. Caloric restriction is, by far, more effective than exercise for net deficit. Thunderf00t did a video on this. https://youtube.com/watch?v=mTABw0EyIWY
Another thing about walking is - while everyone's different, long walks really increased my appetite. While walking 4 miles daily (with some hills) was great in many ways, I slowly gained 20+ pounds over 2 years. Running didn't have same effect on me.
4-5 hours for 20k steps? Because of your heart issues?
I just checked my current stats. I have 15k steps recently from walking to and from my job. And that’s not counting the steps at the gym. I take off my watch since it’s often a hindrance during work outs.
I have a recorded 'hike' (on very steep logging roads) of 7 hours with a total expenditure of 1800 cal active/2500 cal total with an elevation gain of 3000' (the decent was way harder).
Tell me how to burn 3000 calories in 40 minutes and I will be in perfect shape. It's a struggle to burn 400 in 30 minutes (hiit training, very challenging)
