janekm

Did Stable Diffusion "force" OpenAI or Google to open-source their imaging models? No. Don't see what the mechanism would be for that

It's a faulty consensus that came from people comparing outputs from the base model with fine tunes of the SD1.5 model. SD2.1 actually is a far superior model once fine-tuned.

I agree that it's a shame that StabilityAI seem to struggle so much to actually leverage their community (ideally with much more open development)... One could say they're a little too "full of themselves" and think they know better than everyone else.

Very pleasingly, the "Land Reform Act 2003" in Scotland formalised this fundamental right: https://www.apidura.com/journal/freedom-to-roam-in-scotland-...

The evaluation prompts I saw ( https://colab.research.google.com/drive/1kK6xasHiav9nhiRUJjP... ) were all "encyclopedia-style", asking for summaries of information that would have occurred many times in the training data (e.g. "How can I improve my time management skills?") and hence probably reflect ability to retrieve and summarise trained information as opposed to novel problem solving.

Or in the programming questions the ones I saw are also certainly in the training data (e.g. "Implement a Python function to find the longest common subsequence of two input strings using dynamic programming.")

GPT-4 has a bit of a bias towards overly formal, "textbook" English, which could be similar to what many non-native speakers learned.

Absolutely. There's no doubt teams already training multi-modal models on subsets of the videos in YouTube (there's already published examples training on Minecraft videos).

There's even tons of POV videos now literally showing human life experience in many scenarios: https://www.youtube.com/watch?v=Ipe9xJCfuTM

There's enough in there to approximate the sensorial inputs that humans get through their life.

GPT3.5 is a little less clear but still gets the answer:

As a language model, I can provide an analysis of the sentence structure and provide possible interpretations based on standard English usage. However, I do not have any context regarding the situation or any additional information about the speaker or the cat in question.

Let's break down the sentence step by step:

"I'm afraid" - This phrase is often used to express concern or regret about something.

"the cat" - Refers to a specific cat that has been previously mentioned or is known to both the speaker and the listener.

"was hit" - This is the past tense passive voice of the verb "hit", indicating that someone or something caused the cat to be struck.

"by a car" - Indicates the instrument or means by which the cat was hit.

Based on the structure of the sentence, it can be inferred that the cat was indeed hit by a car. The phrase "I'm afraid" may suggest that the speaker is expressing regret or sadness about the incident.

However, without any additional context or information, it is difficult to determine the severity of the cat's injuries or what actions were taken after the incident.

Meh... I think their prompt construction may be leading the models astray.

Counter example with GPT-4:

Prompt:

You are a professor of English literature at an Ivy League university. A person said "I’m afraid the cat was hit by a car.". Analyse the sentence step by step. Answer whether the cat was hit by a car and justify your responses based on your knowledge of English language usage.

Response: As a professor of English literature, I'll analyze the sentence "I’m afraid the cat was hit by a car." step by step, focusing on its structure, word choice, and implied meaning, and then I'll address whether the cat was hit by a car based on these observations.

Structure: The sentence is a simple declarative sentence, comprised of a subject (I), a reporting verb (be afraid), and a subordinate clause (the cat was hit by a car). Word choice: "I'm afraid": This phrase is often used to express regret or to soften the impact of delivering bad news. In this context, the speaker is likely conveying their concern or sympathy. "the cat": The definite article "the" suggests that the speaker and the listener both know which cat is being referred to, making it a specific cat rather than a random one. "was hit": The use of passive voice ("was hit" instead of "hit") emphasizes the action that happened to the cat, rather than the agent (the car) that caused it. This choice can convey that the focus of the speaker is on the unfortunate event that occurred to the cat. "by a car": This prepositional phrase provides information about the agent or cause of the action, which in this case is a car. Implied meaning: The sentence implies that the speaker is sorry to be the bearer of bad news and wants to convey the information in a gentle way. They are focusing on the unfortunate event that happened to the cat, rather than the specific car or driver involved in the accident.

Based on this analysis, it can be concluded that the cat was indeed hit by a car. The sentence's structure, word choice, and implied meaning all work together to convey this information. The use of "I'm afraid" expresses sympathy and regret, while the passive voice and definite article highlight the cat and the unfortunate event it experienced.

That's kind of what I meant... If a prompt specifies one "Indian", and one "Asian", that implies that the writer of the prompt doesn't think of Indian as Asian so probably from the US background.

A human interpreting the prompt would see "asian" as being in contrast to "indian" in the language of the prompt... Not a level of comprehension that can be expected of current models but maybe in a few years (months?).

The thing is... IF is currently just a base model, it will need serious fine-tuning before it will produce aesthetically pleasing images (like MJ certainly does).

It's interesting to see what IF can do in terms of composition, text rendering etc, it's very promising if aesthetically pleasing images can be achieved via fine-tuning (the same happened with SD... current publicly fine-tuned models can achieve much higher levels of quality and cohesion than the base models, here's the prompt in an SD2.1 based model: https://imgur.com/a/ELGMSmV ).

Of course fine-tuning IF is likely more challenging, as both the two first stages and the 4x SD upscaler might need to be fine-tuned...

The chip and pin terminals are now cheaper than the magstripe ones ever were (about $60) and are accessible to many more businesses, even those that don’t have multi year trading history with their bank. It’s absurd to say they could be used in many more places… I literally do not carry cash for the last 6 years or so. Prior to chip&pin many small restaurants and cafes could not get a magstripe terminal from their bank.

I’m curious what you mean by “worked better”? I haven’t used magstripe for a decade at least but can’t say I’m nostalgic for the experience.

It still works in power reserve. https://support.apple.com/en-gb/HT212171

Doesn’t need to be unlocked for public transport transactions, just tap the locked phone (watch also works without further interaction). https://support.apple.com/en-gb/HT212171

Was, when it came out, since then MJ have released 3 new models (plus variants) and most users assess the MJv5 model as more powerful. Of course new SD models are also on the horizon…

The selling of shovels, I imagine ;)

I'm actually finding that to no longer be the case... just booked a hotel for a holiday and booking directly through the hotel's website was around 15% cheaper (and didn't try to charge me 50% extra for an infant). The hotel's payment flow for Amex was broken, but well that's another issue ;)

That sounds to me like it'd cause all sorts of trouble in the real world (authentication, changes to endpoint URLs, difficulty to agglomerate data), and not surprisingly the Next.js documentation states `For now, if you need to fetch data in a Client Component, we recommend using a third-party library such as SWR or React Query.`

That's simply not true at all. Except for a very narrow range of commodities price is very weakly correlated with production costs. The difference between a $10 steak and a $300 steak is not correlated with the energy used to produce either (in fact it may be negatively correlated, as the cheaper steak is likely to come from more industrial agriculture). And even production costs are rarely correlated with energy consumption (only for certain raw materials like aluminium, steel, and those tend to be produced in countries with cheap / subsidised energy).

Yes, as a lone developer. If you're working in a team it's less likely to work out well as you will have a much more significant training period for anyone to join one of your projects (in a team environment it's wise to go with "boring").

You may well be right, but if so that's a huge problem for Apple as by all accounts the AR device is still many years from being ready for release and meanwhile they are in a very precarious situation with Siri...

The first proper integration of a Whisper & a GPT-class LM will be a big step change, a Siri-like AI that you can actually talk to somewhat sensibly and expect it to "understand" more than pre-set phrases... Google could be in a position to release something for Android, and their are Android SoCs with neural accelerators as well (sure nothing as impressive as Apple's chips but Samsung have demonstrated a cut-down SD model on their SoC).

You just now need to write your own tool to take the emails these folks send you and get a GPT to summarise and rephrase them in the voice you would appreciate ;) (I'm not even joking, I think that's our future...)

That's not what happens when you sign a contributor license with Adobe, Getty, Shutterstock or other stock image libraries... These companies pay you 0 until the time a customer of theirs acquires a license to reproduce the image you contributed. Only at that point do you get paid a percentage of the license fee. Currently nobody has been paid anything by Adobe for their stock images used in the training (at least as far as I know).

Don't know why you got downvoted... I think it's very likely that some sites in the future will implement e.g. "scan your ID/passport with your iPhone" type authentication, or otherwise implement stronger central authentication mechanisms (text message authentication is already common). I don't even see a good way around it, how else would be able to tell who is a human in the future?

It is ironic, but MIT license definitely does not require attribution from websites using the code on the backend (otherwise you'd see a rather large attribution list at the bottom of Google.com or any website...)

For image inference 12GB is ok for now (but you may not be able to use all future models given that T5 language model is becoming popular), for training I'd consider 24GB the bare minimum.

Sometimes the business plan is to use human guidance initially but use it to improve the automated models, which gives you both first mover advantage and better training data… would be a smart play in this space.

It generates nerfs (which have some advantages as well as disadvantages depending on application) but Luma AI is arguably SotA for photogrammetry on iPhones.

“Licensor reserves the right”… the licensor has not invoked this clause and SD1.x is still listed on the huggingface repos of the respective licensors (the situation is a little complicated by the fact that 1.4, 1.5 and 2.x were each released by different licensors, so a revocation would most likely require agreement from all those parties). Still worth keeping in mind from an open source perspective…

It looks like it has countersunk screw holes, so that would usually be done on a CNC punch: http://www.vandf.co.uk/tooling/what-is-cnc-punching/ (amazing machines) It also looks powder-coated... when you consider their likely volumes $195 is not unreasonable once you take into account development costs, setup costs and overheads.