HNHacker News
TopNewBestAskShowJobs

sgt101

7,940 karma · joined January 24, 2012

nothing to see here, move along, move along.
submissionscomments
sgt101··on Britain is building one of the world’s most expensive railways
On the other hand there's a big part of the UK that doesn't want to work or support itself and does want to confiscate other peoples money in order to continue to sit in social housing with social wellfare and a plethora of other social services.

We can have a libertarian country.

Or we can have a social democratic one.

We can't have both.

sgt101··on Britain is building one of the world’s most expensive railways
There's no mystery, it's just very orderly corruption.
sgt101··on A $12k Surgery to Change Eye Color Is Surging in Popularity
When I was a young fella I had a thing for folks with green eyes - I ended up marrying one!

So - don't knock it!

sgt101··on The Battle Line at Louvain
Very interesting to note that Oxford wasn't extensively bombed in ww2 despite being a manufacturing hub. I wonder what role this incident had in that choice.
sgt101··on Congress hears testimony on aliens the government hides
Especially ones not written in crayon.
sgt101··on The Onion buys Infowars
That's what I guessed from the story - I suppose that there would be a lot of reticence about buying this site from a lot of corporate actors, but maybe there are a lot of crazy people who could have bought it if the price was right. So, this way it is for sure a "dead" property.
sgt101··on The Onion buys Infowars
I think it's all the assets of the website.

I believe that the current state of play is that Jones has to pay $1.1bn damages even post bankruptcy so maybe any future successes will lead to money for the Sandyhook families. I certainly hope so.

Ironically he may live longer to earn more for them - he'll never be able to afford a cigar again.

sgt101··on Experts testify at UFO hearing in Congress
well - all these people believe in a big spirit in the sky, so it's not a huge stretch
sgt101··on NASA remains silent on why crew went to hospital after dragon splashdown
But the moon is 1/9th the mass of Mars (approx)

The moon has an average orbital velocity of 1km per sec (approx)

Mars is 24km per sec (approx) Terminal velocity is 600ms, on earth it's 54ms

So we are talking about something that is an order of magnitude more difficult. I think if retro rockets are to be used then far more fuel will be needed. But that means far more mass, which means more fuel - how much fuel would be needed to slow something like Starship for example?

sgt101··on NASA remains silent on why crew went to hospital after dragon splashdown
I looked into manned Mars missions in a fit of enthusiasm 30 years or so, and I concluded that we couldn't do it because landing a human + life support on Mars requires deceleration technology we just don't have. Basically the speed of a Martian orbiter needs to be reduced to 0 or so to avoid splatting the human, but Mars' atmosphere is so thin that parachutes just don't provide the required break.

Has that now changed or is it still just fantasy?

sgt101··on The Weeds Are Winning
Yeah - be a little careful there.

I've been playing with llama-3.2vision-11b and it can't count more than 4 dots.

sgt101··on Ollama 0.4 is released with support for Meta's Llama 3.2 Vision models locally
I've tried with some ppt images rather than Clevr ones and it does much better. It can count circles and triangles and differentiates between them quite well. It can recognise the colours of the objects as well.

I think that the faux 3d of clevr images is too much for the model, it's interesting because much smaller pre-transformer specialist models were very good at clevr.

sgt101··on Ollama 0.4 is released with support for Meta's Llama 3.2 Vision models locally
I tested the small model with a few images from Clevr. On first blush I am afraid it didn't do very well at all, it got object counts totally wrong and struggled to identify shapes and colours.

Still, it seems to understand what's in the images in general (cones and spheres and cubes), and the fact that it runs on my mac book at all is basically amazing.

sgt101··on The deep learning boom caught almost everyone by surprise
I'm so out of things ! What do you mean manually computed derivatives?
sgt101··on DeepMind debuts watermarks for AI-generated text
Ok, so what protects you from these folks? What positive measure can be suggested here - that is better than the measures I suggest and subsumes them?
sgt101··on The deep learning boom caught almost everyone by surprise
mid-60's had neural nets.

mid-90's had LeCun telling everyone that big neural nets were the future.

sgt101··on The deep learning boom caught almost everyone by surprise
It's not quite so - we couldn't handle it, and we didn't have it, so it was a bit of a none question.

I started with ML in 1994, I was in a small poor lab - so we didn't have state of the art hardware. On the other hand I think my experience is fairly representative. We worked with data sets on spark workstations that were stored in flat files and had thousands or sometimes tens of thousands of instances. We had problems keeping our data sets on the machines and often archived them to tape.

Data came from very deliberate acquisition processes. For example I remember going to a field exercise with a particular device and directing it's use over a period of days in order to collect the data that would be needed for a machine learning project.

Sometime in the 2000's data started to be generated and collected as "exhaust" from various processes. People and organisations became instrumented in the sense that their daily activities were necessarily captured digitally. For a time this data was latent, people didn't really think about using it in the way that we think about it now, but by about 2010 it was obvious that not only was this data available but we had the processing and data systems to use it effectively.

sgt101··on DeepMind debuts watermarks for AI-generated text
Here's an example of why I think this would work.

GDPR.

How many breaches of privacy by large orgnaizations occur in the EU? When they occur, what happens?

On the other hand - what's the story in the USA?

Alternatively what would have happened if we simply said "data privacy cannot be maintained, no laws will help"?

sgt101··on DeepMind debuts watermarks for AI-generated text
well - trusted in the sense that the unwatermarked text was human generated ;o)
sgt101··on DeepMind debuts watermarks for AI-generated text
I think this misses a key point.

If there were a law that AI generated text should be watermarked then major corporations would take pains to apply the watermark, because if they didn't then they would be exposed to regulatory and reputational problems.

Watermarking the text would enable people training models to avoid it, and it would allow search engines to determine not to rely on it (if that was the search engine preference).

It would not mean that all text not watermarked was human generated, but it would mean that all text not watermarked and provided by institutional actors could be trusted.

sgt101··on M4 MacBook Pro
You have another one with a network gateway to provide hot failover?

Right?

sgt101··on New iMac with M4
Imagine you are running a digital design shop and you want to determine two things:

- the software that your users are running is licensed. - the media that is produced is stored on a central resource that's tightly managed and comprehensively backed up.

If so, maybe having client machines with minimal disk space is what you want?

sgt101··on Use Prolog to improve LLM's reasoning
Building on this idea people have grounded LLM generated reasoning logic with perceptual information from other networks : https://web.stanford.edu/~joycj/projects/left_neurips_2023
sgt101··on Why don't we use awnings anymore (2022)
Definitely on my mind in the UK. We have maybe 10 days a year when cooling is important, but having sun shades would be really nice for perhaps 50 days a year - so makes more sense than an AC solution here I think.

Also they look nice!

sgt101··on Machine learning and information theory concepts towards an AI Mathematician
I love that "system 1" and "system 2" are tossed out as fundamental concepts of mind when they're basically labels for some semi-random ideas based on bullshit studies in a pop sci fluff book. Is there any actual science that ties these ideas to mathematical reasoning?

Then we have:

>> By analogy, we can now state the main hypothesis proposed in this paper: a crucial component of the usefulness of a new proven theorem t (in the context of previous theorems T(S)) is how efficiently T(S)∪{t} compresses the set of all provable mathematical statements M. That is, T (S) ∪ {t} is a good compression of M if many provable statements, beyond those in S, can be derived from T (S) ∪ {t}, say using at most k derivation steps.

which says "Occam's Razor, that's good isn't it?" in more words.

sgt101··on [dead]
I've got my tin foil hat on today! There is a story about how one of Mike Lynchs closest business associates (ex ceo of Darktrace) as been made a minister of state in the UK.

Pretty clear signal that she's not to be touched. I don't think hurricanes are engineered, or elections get stolen.. but I do think that legal authority in the USA and sphere is being very strongly tested.

sgt101··on One Genius' Lonely Crusade to Teach a Computer Common Sense (2016)
I thought that the Apollo system was barely able to execute on the available hardware, so there was no redundancy.

I think that the space shuttle did feature this kind of system though.

sgt101··on Alternatives to cosine similarity
May be the author could test with realistic data set sizes. Recently I've worked with 26M and 300M data sets. Perhaps something along those lines.
sgt101··on Is the attack helicopter dead?
What a pity that Brimstone isn't integrated with the Apache.
sgt101··on William Cowper and the Age of the Earth [pdf] (2019)
My uncle believed in bible chronology. He believed it so much that he learned to program computers (to calculate and explore patterns in dates) and then built a series of web sites culminating in https://www.biblechronology.com/

When he died he had assembled a team of people who worked on it. The site is now suffering from a bit of rot (the videos don't serve anymore) but he has been dead for six years so I suppose that's no surprise. I don't know if his "gang" still maintain it, but someone must be paying for the domain and hosting I suppose.

The time and energy he committed to this are astonishing to me, he was a talented man - he had the option to be a partner at Arthur Anderson (progenitor of Accenture) before becoming a CFO at a series of small banks and building societies. His career was firmly on the up, he worked for one of the precursors of what became RBS in the 1980's. If things had gone differently perhaps he would have muzzled "Fred the Shred" and we'd all be the richer, especially his kids. But, at some point a conviction and faith gripped him and he gave up everything for his bible project. He died penniless.

I wonder, because I don't understand, because I just see numerology and over interpretation, does it mean he was wrong? I think so, but perhaps that's just my faith talking.

← PreviousPage 14 of 34Next →