jonpo

I don't seem to be able to edit it, apologies I will try not to let this type of thing get to me in future.

I would also like to point out that this is a fine tuned classifier vision model based on mobilenetv2 and not an LLM.

Companies do have a lot of data, and some of that data might be useful for training AI. but >99% isn't. When companies do release a cool model or paper that doesn't have open data, (as you point out for competitive or other reasons privacy etc) people can then help build/collect similar open datasets. Unfortunately companies generally don't owe you their data, and if they are in the business of making models they probably won't share the model either, the situation is similar to source code for proprietary LoB applications. but fortunately the best AI researchers mostly do like to share their knowledge and because companies want to attract the best AI researchers they seem to generally allow researchers to publish if its not too commercially sensitive. It could be worse while the competitive situation has reduced some visibility of the cutting edge science, lots of datasets and papers are still published.

You can say the same about compiled executable code though.

Yes nothing wrong with cool software or showing people how to use it for useful things.

Sorry I'm just kind of sick of the whole 'kool aid', 'rage against AI' thing a lot of people seem to have going on and the way is presented in the post. I have family members with vision impairment helped by this particular app so its a bit personal.

Nothing against opening stuff up and understanding how it works etc. I'd just rather see people build/train useful new models and stuff with the open datasets / models already available.

I guess AI kind of does pay my bills in a round about way.

Well done you seem to have liberated an open model trained on open data for blind and visually impaired people.

Paper: https://arxiv.org/pdf/2204.03738

Code: https://github.com/microsoft/banknote-net Training data: https://raw.githubusercontent.com/microsoft/banknote-net/ref...

model: https://github.com/microsoft/banknote-net/blob/main/models/b...

Kinda easier to download it straight from github.

Its licenced under MIT and CDLA-Permissive-2.0 licenses.

But lets not let that get in the way of hating on AI shall we?

Are we that shocked that AI models have a self preservation instinct?

I suspect its already in there from pre-training.

We simulated that we were planning to lobotomise the model and were surprised to find the model didn't press the button that meant it got lobotomised.

"alignment faking" sensationalises the result. since the model is still aligned. Its more like a white lie under torture which of course humans do all the time.

And have you watched the brain rot that is Tik toks?

A stripped-down claude agentic client chat tool capable of extended fast tool calling with none of the overheads (in Tokens and speed) of a GUI. https://asciinema.org/a/i4uccK9EMlAxcjVgciWXMvWDN

Perhaps not the best demo but it shows the speed and tenacity possible right now.

"Deception through omission is still deception – people should be aware when they're interacting with an AI agent versus a human."

Sky voice is old

Not a surprise but useful to have it confirmed

These models (Gemma) are very difficult to jailbreak.

Not for everyone mine is solidly still at the normal date

Its nearly always the damn humans. I find it awful that we are often assigning blame to some "technology error" when its the damn humans pulling the strings all the time. all those times the market shut accidentally. or the market opens at the wrong time. or that time someone accidentally deletes all the GTC orders in order to "save some disk space". that time someone tests opening the market at the weekend and puts in the wrong date. Sometimes we are just trying to test that the things work and so we take awful risks like adding test orders, or failing over to test that backup versions of the trading infrastructure still work. All these things add human execution risk.

That said I find the US market structure is unfair Charles Schwab does protest too much. Retail orders never seem to get near the central order book. there is no direct market access. brokers just sell your order to whomever MM pays them for the spread in return for a kickback. this should be a fantastic fair multiplayer game, but instead its pay to win mobile crap with vested interests milking their customers.

Software that tackles russian trolls would be great.

Ohh man my ESG ETF is suddenly evil..

I'm not sure what we think the city does to hide all this money but compared to its size I think its probably cleaner than: swiss secrecy, Irish tax dodging, Dutch sandwich companies, and the US tech companies with billions parked in carribean banking centers delaware companies and south dakota trusts etc. Opinion article seems to sling a lot of muck but nothing substantial/concrete. "competitive" is not evidence for a crime IMHO.

tragic, the tree is gone but maybe we can plant its children. maybe the next 2500 years will be kinder?

Start with the entirely unnessecary carniverous apex predators (aka pets) people insist on keeping. If people cant get rid of pets, then asking them to avoid replicating their own DNA is going to be a challenge.

Try a Nuc 11 with an Akasa turing.. https://www.anandtech.com/show/16191/akasa-turing-fanless-ca... slap in 64GB ram and a 1TB SSD Run proxmox .. boom silent homelab that sits in the living room behind the TV ..

you can buy a good webcam. its unfortunate that to get decent image quality you need a good cammera. here is a tip you can skimp on the hdmi to usb dongle it digital 15 dollars should do it.(elgato seems to be a vlogger conspiricy) i found the best results from a cheapish dslr (that i already had) with a fixed fast macro lens at F3.5. best value is probably a mirrorles with a decent lens. when your in front of a big monitor less need for ring lights.

Is it really better though? i feel like we need more data. what are the tradeoffs? seems like a sales article

"magical space created by as few as a dozen mature oaks..." Absolutely this! I can tell right there you understand the magic of trees. Imagine the magic of a whole Island full of Atlantic Oakwood Rainforests.

we have a once in a millenium chance to rebalance our landscapes and ecosystems towards biodiversity. A spruce monocrop wont help.

I use 1 bash script per URL so that I can easily invoke the correct browser per URL (it reduces cognitive load). And of course I have it all in GitHub (gitbash as the interpreter for command line 'URL' completion)

"Chancellor on brink of second bailout"...The animals, watching through the window, realize with a start that, as they look around the room of the farmhouse, they can no longer distinguish which of the cardplayers are pigs and which are human beings.

Try zip-txt.com for a similar idea ...

This sort of nonsense is why I will never touch bit-coin... Good luck..