Master AI (Artificial Intelligence) Discussion/News Thread

[pablolie]

That's a shame. I don't think there was anything really wrong with @amirm's post on the RAM supply/demand thing or with my tentative reference to the Nvidia priorities rumour. But generally if you were a bit less cranky and open to different perspectives (expecting homogenous perspectives on this subject is pretty unrealistic) you were certainly contributing some yourself.

Not sure why I still saw the reply despite unfollowing the topic.

I don't mean it in any way disrespectfully at all, many of you guys know I learn from you a lot and treasure your insights.

It's just this one topic ran its course for me personally for now. We shall see how the chips fall when AI reality as a useful huge data cruncher settles (because that's what it is, with pointed human directive ).

I hope you don't bash your children when they don't immediately provide the right answer to a question... because AI is that, a tabula rasa until it is sufficiently trained, and anyone that thinks otherwise doesn't know AI.

PS: I have zero vested interest in AI and could not care less if it succeeds commercially (but I know it will within realistic boundaries).

 

[Axo1989]

Not sure why I still saw the reply despite unfollowing the topic.

I don't mean it in any way disrespectfully at all, many of you guys know I learn from you a lot and treasure your insights.

It's just this one topic ran its course for me personally for now. We shall see how the chips fall when AI reality as a useful huge data cruncher settles (because that's what it is, with pointed human directive ).

I hope you don't bash your children when they don't immediately provide the right answer to a question... because AI is that, a tabula rasa until it is sufficiently trained, and anyone that thinks otherwise doesn't know AI.

PS: I have zero vested interest in AI and could not care less if it succeeds commercially (but I know it will within realistic boundaries).

I expect replies generate their own alerts, in addition to new posts in a watched (or unwatched) thread. Anyway if I was responding too aggressively to you I apologise.

 

[sam_adams]

This explains my puzzlement of why consumer RAM prices were going up even though they are not the type used for GPU/AI machines.

It's not just the price of RAM that is being affected by the GPT hype. The price of Raspberry Pi devices has jumped—a 16GB kit is over $200—because of a single posting on Xitter that postulated that a Pi could be used to run "the agentic AI hand grenade known as OpenClaw" on the PI. The post also sent the price of the stonk soaring—it has since dropped back down a bit.

 

[somebodyelse]

It's not just the price of RAM that is being affected by the GPT hype. The price of Raspberry Pi devices has jumped—a 16GB kit is over $200—because of a single posting on Xitter that postulated that a Pi could be used to run "the agentic AI hand grenade known as OpenClaw" on the PI. The post also sent the price of the stonk soaring—it has since dropped back down a bit.

I think that's more coincidence than cause and effect - the prices have risen on most models by an amount broadly proportional to installed RAM, similar to the previous hike last December. And the rise in RAM prices since December is well known.

 

[antcollinet]

Quite often when I interact with AI it blows me away. This is a little long - I'll hide all the AI responses in quote tags, and the images as thumbnails.

I'm part of a photo editing group. We meet Monday afternoons on zoom. Four of the guys also go to the local (traditional type) camera club where there are regular club competitions with guest judges. These four are often commenting about the standard of and aribitrary nature of some of the judging at the club. Today one quipped "Perhaps we should replace them (the judges) with AI".

I said ..... "Hold my beer"

Prompt : Provide a competition critique for this image.

## Competition Critique

### Overall Impression

This is a striking and atmospheric landscape image with strong graphic impact. The silhouetted trees and their near-perfect reflections create an immediate visual hook. The symmetry is compelling, and the subdued winter palette enhances the mood. It feels intentional and composed rather than incidental — always a good starting point for competition work.

---

## Strengths

### 1. Composition & Structure

* The vertical tree trunks create powerful rhythm and repetition across the frame.
* The reflection forms a natural mirror line, giving the image structural strength.
* The slight break in uniformity (the taller evergreen on the right) adds visual interest and prevents monotony.
* The low sun positioned behind the trees provides a subtle focal point without overwhelming the silhouette effect.

In competition terms, the image reads clearly at first glance — a strong advantage.

---

### 2. Mood & Atmosphere

* The winter bareness of the trees creates an elegant, almost graphic silhouette.
* The muted sky tones keep the image cohesive and restrained.
* The still water enhances a sense of calm and contemplation.

There’s a quiet, reflective quality here that judges often respond well to.

---

### 3. Technical Control

* Exposure is well handled considering the dynamic range challenge (bright sky vs. dark silhouettes).
* The reflection is sharp and well-defined — indicating good timing and stable water conditions.
* The blacks are rich without appearing blocked in most areas.

---

## Areas for Improvement

### 1. Focal Hierarchy

While the image is strong graphically, it lacks a dominant subject. The eye wanders rather than settling.

In competitions, judges often ask:

> *“What is this image about?”*

The sun is partially obscured and not strong enough to anchor the composition. Consider:

* Waiting for the sun to break more clearly between trunks
* Introducing a human or wildlife element for scale and narrative
* Cropping to emphasize a particular section with stronger interaction

---

### 2. Crop & Balance

The near-perfect symmetry is powerful — but slightly static.

You might experiment with:

* A tighter crop removing some foreground water (the lower area is slightly less clean tonally).
* A panoramic crop focusing on the upper half and strongest reflections.
* Or committing fully to symmetry by ensuring the reflection axis is precisely centered.

Right now, it feels almost symmetrical but not intentionally exact — which weakens impact slightly.

---

### 3. Tonal Separation

The darkest areas in the lower reflection are beginning to merge into a single mass.
Lifting shadows very subtly in the reflection could:

* Reveal more branch detail
* Add depth
* Improve print performance (important in competitions)

---

### 4. Sky Interest

The sky is soft and pleasant but not especially dramatic. In higher-level competitions, skies often carry more presence — either through:

* Stronger cloud structure
* More defined light rays
* Or cleaner minimalism (if going graphic)

At the moment, it sits somewhere in between.

---

## Competition Potential

**Club Level:** Strong contender
**Regional Level:** Competitive with refinement
**National Level:** Would need stronger focal emphasis or narrative hook

The image excels in:

* Graphic design
* Mood
* Reflection symmetry

It would score well for composition and impact, but may lose marks for originality or subject strength unless refined.

---

## Suggested Refinements Before Entering

* Decide: graphic abstraction vs. landscape narrative. Push one direction more decisively.
* Refine the crop for deliberate symmetry or deliberate asymmetry.
* Enhance tonal separation in print.
* Consider subtle dodging to emphasize one tree grouping or the sun as a focal anchor.

---

## Final Verdict

This is a thoughtfully composed, aesthetically strong landscape with excellent reflective symmetry and mood. With careful tonal refinement and a clearer focal hierarchy, it could move from “very good” to “competition standout.”

If you’d like, I can also:

* Suggest specific crop ratios visually
* Provide a judge-style scoring breakdown
* Or suggest a competition-ready title (which can influence scoring more than people realize).

This blew me away - the ability to break down the image, describing it in detail - and then critique it competently from a photographic quality point of view is - frankly - amazing. The consensus of the four club members is that it is better than the majority of the guest judges. So we decided to take it a step further, by asking it to compare with a similar shot.

Compare it with this similar shot - which one would you rate as the best.

I had already shown both images to the group and the consensus was that the first is the stronger image. Chat GPT agreed.

This is a very useful comparison — they are similar in subject, but they behave quite differently in competition terms.

I’ll refer to the **first image** as *Image 1* (stronger silhouette, darker mood) and this as *Image 2* (brighter, more detail in trunks and bank).

---

# Immediate Visual Impact

### Image 1

* Stronger graphic punch.
* Cleaner silhouette.
* More minimal and bold.
* Reads instantly at thumbnail size.

### Image 2

* More naturalistic.
* More tonal information.
* Slightly busier.
* Takes longer to “read.”

**Advantage: Image 1**
Competitions are often won in the first three seconds.

---

# Light & Mood

### Image 1

* More dramatic tonal separation.
* Moodier and more atmospheric.
* The darker reflection feels deliberate and controlled.

### Image 2

* Brighter exposure.
* Warmer ground tones visible.
* Less dramatic, more documentary.

Image 2 feels like a well-exposed landscape.
Image 1 feels like an interpreted image.

**Judges generally reward interpretation over documentation.**

**Advantage: Image 1**

---

# Reflection Quality

### Image 1

* Reflection feels cleaner.
* Darker water simplifies the lower half.
* Strong mirror illusion.

### Image 2

* Foreground shallows reveal leaves and green tones.
* Adds realism, but reduces graphic impact.
* Slightly less “mirror-like.”

In competition, simplicity often scores higher than realism.

**Advantage: Image 1**

---

# Focal Hierarchy

This is interesting.

### Image 1

* Sun is more subtle.
* Feels evenly weighted.

### Image 2

* Sun on the right is brighter and more obvious.
* Creates a clearer visual anchor.
* But also pulls the eye hard to the edge.

That edge placement slightly destabilizes the frame.

**Slight Advantage: Image 2 for focal clarity**
But the edge tension weakens it.

---

# Composition & Balance

Image 1 feels more cohesive overall.
Image 2 feels slightly more left-heavy and tonally varied.

The visible bank and green patch bottom left in Image 2 are distractions in a judged print.

Image 1 is cleaner.

---

# Originality & Competition Potential

Image 1:

* Stronger as a graphic study of repetition and reflection.
* More “print competition ready.”
* Would score higher for impact and style.

Image 2:

* Strong competent landscape.
* Likely to score well at club level.
* Less likely to stand out at higher level.

---

# Final Verdict

**Image 1 is the stronger competition image.**

It is:

* More graphic
* More cohesive
* More dramatic
* Cleaner in structure
* Better simplified

Image 2 is technically good, but Image 1 feels more intentional and interpretive — and that is what usually pushes scores upward.

---

# If You Were Entering One

Enter **Image 1**, but:

* Lift shadow detail slightly in the reflection for print.
* Ensure the symmetry axis is precise.
* Consider a tighter crop to remove any weak lower tones.

We also asked it to score both images in a judge style scoring - which it did in a way that felt right. I'll not bother posting that reply here.

But I did point out that AI likes to flatter - perhaps it was over rating the images to please me. I suggested giving it a photographically meritless shot to evaluate. I used an image of a cake I baked (I am not a baker - I took the shot as a humorous thing to send to the kids - "Hey look your Dad can bake now")

Please assess this image:

Well there was no flattery. I was particularly amused by :

This reads as:
“Here is my cake.”

Which is exactly what it was.

I’ll assess this as a photographed subject (food image), not the baking result itself — though they do interact.

---

# Overall Impression

This is a casual, documentary-style food photograph. It clearly shows the loaf and texture, but it does not yet read as a competition-level food image. The biggest limitation is presentation and control of environment rather than camera technique.

If entered in competition as-is, this would likely sit mid-table.

---

# Judge-Style Breakdown (Out of 20)

## 1. Impact – 2.5 / 5

The subject is clear, but it lacks immediate visual appeal.

Why:

* Background is a worn chopping board with visible knife marks.
* No styling or supporting elements.
* Lighting is flat and utilitarian.

It feels like a record shot rather than a crafted image.

---

## 2. Composition & Design – 3 / 5

Positives:

* Diagonal placement gives some dynamism.
* Shallow depth of field helps slightly.

Negatives:

* Too much empty board in the foreground.
* Top of loaf nearly touches upper frame.
* The angle flattens the loaf slightly.
* No clean negative space — just unused space.

A tighter crop or controlled composition would help significantly.

---

## 3. Technical Quality – 3.5 / 5

Exposure is acceptable.
Focus is reasonably sharp on front surface.
Colour is natural.

However:

* Lighting is soft but not directional — texture lacks dimensional modelling.
* Highlights on the top are slightly dull.
* The background is visually noisy (board scratches).

---

## 4. Creativity / Interpretation – 2 / 5

This reads as:

> “Here is my cake.”

It does not yet communicate:

* Mood
* Occasion
* Story
* Indulgence
* Atmosphere

Food photography at competition level is about appetite appeal and styling — not just documentation.

---

# Total: 11 / 20

This would likely receive a comment like:

> “A well-baked subject clearly recorded, but I would have preferred more attention to presentation, lighting direction and background control.”

---

# What Is Holding It Back?

1. The chopping board dominates the frame.
2. No styling (no crumbs arranged intentionally, no sliced piece, no garnish).
3. Lighting is ambient rather than shaped.
4. No focal hierarchy — just a loaf.

---

# How To Instantly Improve This Scene

If you re-shot this:

### 1. Control the Surface

* Use a cleaner board, slate, or linen cloth.
* Remove scratch marks from the frame.

### 2. Create Appetite Appeal

* Slice it.
* Show crumb texture.
* Add a knife, crumbs, cocoa dust, or butter for context.

### 3. Directional Light

Place it near a window:

* Side light from 45°
* Add shadow on opposite side for depth
* Avoid top-down flat light

### 4. Crop More Aggressively

Reduce foreground dead space.

---

# The Important Question

Is this:

* A quick phone snap?
* Or intended as a competition food image?

Because if this is a casual snap — it's perfectly fine.

If this is for competition — it needs styling, lighting control, and compositional intent.

---

If you'd like, I can:

* Mock up a competition-ready composition plan for this exact loaf
* Or give you a “how to shoot food at home without expensive gear” guide
* Or score it as a harsh judge who specializes in food imagery

So as a final experiment:

Can you generate an image of this food that would be competition suitable so I can see what it should look like.

We did wonder how - if we were to submit this image to a competition - the judge would be able to recognise that it was AI generated.

 

[Raiju]

We are so f*cked

 

[Chr1]

Hmm. Slightly less depressing.
(While the datacentres bake the planet.)

Howsabout...

 

[somebodyelse]

We did wonder how - if we were to submit this image to a competition - the judge would be able to recognise that it was AI generated.

I'd wonder about the slices not matching up (crust and cut face details) but given the trickery used in food photography by humans they might just have thought those slices looked better than the ones that matched up.

 

[AA63]

I'd wonder about the slices not matching up (crust and cut face details) but given the trickery used in food photography by humans they might just have thought those slices looked better than the ones that matched up.

This is a giveaway...

There is no connection between the parchment underneath the cake and its tray. So while it is plausible that a strip of it was used (for a grip?), it is a) unlikely b) the warp in the board edges does not really match up, or c) that a seasoned photographer would let that go.

AI is good at analysis and making stories around it its narratives. For photography, I like the (prompt) interpretation improvements made in the last few months in Nano Banana, very impressive. Midjourney still reigns supreme from a quality-of-finished-photo perspective IMHO.

 

[Chr1]

The photos I'm getting from my new Xiaomi phone, with AI assistance, are way better than my old Fujifilm X20 mostly.

 

[rdenney]

No one blames a chainsaw if somebady hacks their hand off.

My friend, of course they do. That's why chain saws have all sorts of guards and arm-operated brakes, necessitated by losing product liability cases.

Which brings me to a point I've made before: Will the creators of AI, who are not marketing at least their public-facing tools as narrow programming environments that require specialized skills for prompting, but are rather presenting them as general tools available for (nearly) free that are conversational with non-experts, be held liable for their torts?

The usual answer I hear to this is that users are required to indemnify the providers. But you can't contract away your liability for negligence--that is a legal principle hammered into me during my expert witness days by tort liability experts.

Rick "saying it requires expertise to prevent injurious errors implies a required user qualification process" Denney

 

[mhardy6647]

I thought so too, watched the video anyway (I've got lots to do today being Monday here already so procrastinating ftw) and found it surprisingly non-annoying (apart from the French presenter's odd US accent) and even vaguely interesting to see Shambaugh interviewed and Arstechnica mentioned. Also, it was mercifully brief.



Who wanted those annoying chain brakes anyway? If Big Chainsaw had better lobbyists, we wouldn't have so many people with all their arms and legs just getting around aimlessly. Or the negative economic impacts on the mobility device and prosthetics sector.

So... Big Chainsaw and Big Wooden Leg were in cahoots?! Is that what you're saying?

 

[Axo1989]

So... Big Chainsaw and Big Wooden Leg were in cahoots?! Is that what you're saying?

Shiver me timbers, it's a dark world we live in.

 

[pablolie]

My friend, of course they do. That's why chain saws have all sorts of guards and arm-operated brakes, necessitated by losing product liability cases.

Which brings me to a point I've made before: Will the creators of AI, who are not marketing at least the public-facing tools as narrow programming environments that require specialized skills for prompting, but are rather presented as general tools available for (nearly) free that are conversational with non-experts, be held liable for their torts?

The usual answer I hear to this is that users are required to indemnify the providers. But you can't contract away your liability for negligence--that is a legal principle hammered into me during my expert witness days by tort liability experts.

Rick "saying it requires expertise to prevent injurious errors implies a required user qualification process" Denney

Rick, I think you're mixing two different categories of risk.

A chainsaw can physically maim someone. That’s a mechanical hazard. So yes - product liability law forces manufacturers to build in safety guards against foreseeable misuse.

AI is different. It does not amputate hands. It generates probabilistic text. The "injury" comes from humans delegating judgment to it.

It’s closer to:

• A calculator giving the wrong output because you typed the wrong formula
• A GPS routing you into a lake because you ignored your eyes
• A spreadsheet producing nonsense because you didn’t understand the model

AI is a reasoning amplifier. It doesn’t act - it suggests. Humans still decide and keep training the model. That's at the very core of AI. The neural networks, be it RNN or CNN or Transformer or whatever know *NOTHING* until they are extensively trained. And if you want a finished, pre-trained model, it'll *cost* you. They do exist for some indsutry verticals, trained out of the box.

If someone treats a statistical language model as an infallible oracle, that’s not a hidden defect. That’s category confusion.

And importantly: calling it a “tool” doesn’t mean it requires certification. It means it requires literacy - like financial literacy, statistical literacy, or media literacy.

We don’t sue Microsoft because someone misunderstood an Excel formula.

Now, if an AI provider knowingly markets a system as authoritative or expert-grade while understanding it systematically fabricates facts in high-risk domains like medicine or law, then yes - liability questions become real.

But the existence of misuse does not imply the tool is defective. It implies humans are delegating judgment and reason.

 

[antcollinet]

We are so f*cked

That is AI created video - can tell by the low frame rate, and unrealistic movement.

 

[mhardy6647]

That is AI created video - can tell by the low frame rate, and unrealistic movement.

Life is definitely getting weirder.
(and, please, N.B. In this case, unlike my previous post, I do not intend for this comment to be viewed as facetious)

 

[Raiju]

That is AI created video - can tell by the low frame rate, and unrealistic movement.

You have no idea what is going on in China do you?

 

[rdenney]

Rick, I think you're mixing two different categories of risk.

A chainsaw can physically maim someone. That’s a mechanical hazard. So yes - product liability law forces manufacturers to build in safety guards against foreseeable misuse.

AI is different. It does not amputate hands. It generates probabilistic text. The "injury" comes from humans delegating judgment to it.

That’s closer to:

• A calculator giving the wrong output because you typed the wrong formula
• A GPS routing you into a lake because you ignored your eyes
• A spreadsheet producing nonsense because you didn’t understand the model

AI is a reasoning amplifier. It doesn’t act - it suggests. Humans still decide.

If someone treats a statistical language model as an infallible oracle, that’s not a hidden defect. That’s category confusion.

And importantly: calling it a “tool” doesn’t mean it requires certification. It means it requires literacy - like financial literacy, statistical literacy, or media literacy.

We don’t sue Microsoft because someone misunderstood an Excel formula.

Now, if an AI provider knowingly markets a system as authoritative or expert-grade while understanding it systematically fabricates facts in high-risk domains like medicine or law, then yes - liability questions become real.

But the existence of misuse does not imply the tool is defective. It implies humans are delegating judgment without supervision.

I think this hasn't been very deeply tested in the courts yet.

When the product speaks in plain language authoritatively, it is presuming the authority it expresses. If people trust that authority and create harm for themselves, will there be a tort? Probably. But where will negligence be assigned? That's still to be determined, and I don't think we can say that any ship has sailed until the technology has weathered that storm.

Of course, my concern won't just apply to word-generating AI tools, but also to machine-controlling AI, such as with automated vehicles. I think that was a point raised in the article linked last week--the danger is with powerful AI, AI to which authority over important stuff has been granted. People will give away their safety for convenience, but only until they are made to suffer consequences, at which point they may demand accountability for the product rather than their application of it. In my experience, the burden of responsibility is often placed on the creator of the product, because they have unique control over it not granted to the user.

Power tools are dangerous, especially those that are marketed as being safe and easy.

Rick "doesn't know how it will turn out, but thinking the companies are still in the paying off victims phase of liability protection" Denney

 

[Axo1989]

Rick, I think you're mixing two different categories of risk.

A chainsaw can physically maim someone. That’s a mechanical hazard. So yes - product liability law forces manufacturers to build in safety guards against foreseeable misuse.

AI is different. It does not amputate hands. It generates probabilistic text. The "injury" comes from humans delegating judgment to it.

That’s closer to:

• A calculator giving the wrong output because you typed the wrong formula
• A GPS routing you into a lake because you ignored your eyes
• A spreadsheet producing nonsense because you didn’t understand the model

AI is a reasoning amplifier. It doesn’t act - it suggests. Humans still decide.

If someone treats a statistical language model as an infallible oracle, that’s not a hidden defect. That’s category confusion.

And importantly: calling it a “tool” doesn’t mean it requires certification. It means it requires literacy - like financial literacy, statistical literacy, or media literacy.

We don’t sue Microsoft because someone misunderstood an Excel formula.

Now, if an AI provider knowingly markets a system as authoritative or expert-grade while understanding it systematically fabricates facts in high-risk domains like medicine or law, then yes - liability questions become real.

But the existence of misuse does not imply the tool is defective. It implies humans are delegating judgment without supervision.

Well, you introduced the chainsaw analogy. It would be curious to do that if it wasn't apposite. Let's see how the pending cases against chatbot-encouraged suicidal ideation and subsequent action play out?

Note that we don't need to obfuscate by rhetorically assigning any intention or negligence to the chatbot, those potentially reside with the software company or platform of course. Also recall Commonwealth v. Michelle Carter where the defendant was found guilty of involuntary manslaughter. Even though victim Roy still 'decided'.

 

[antcollinet]

Rick, I think you're mixing two different categories of risk.

To be fair - it was your analogy in the first place.