imdsm

20 ms·

301.
Overton Window(en.wikipedia.org)
1 points by imdsm 3y ago|0 comments
302.
Double-Loop Learning(en.wikipedia.org)
3 points by imdsm 3y ago|0 comments
303.
by imdsm 4y ago
That's a great idea, thanks easygenes!
304.
by imdsm 4y ago
FYI, I forgot to make the data persistent[1] so the first five images were sadly lost just now as I made a few updates to the repo. Apologies for this. [1] https://github.com/adamkdean/gpt-unicorn/commit/08e40
305.
by imdsm 4y ago
I don't think any individual sample will tell us much. I think we'll need to review this in 3 months, 6 months, 12 months etc, and look for patterns of changes. Ultimately, this is just a fun side project, something to leave runni
306.
by imdsm 4y ago
Perhaps. I think right now the sample of 5 days isn't anything we can draw conclusions from, but given 365 days, with a number of model updates in between, we may see more interesting patterns.
307.
by imdsm 4y ago
Currently it shows Today (which is always today) and then the image name, `image-2023-04-13` etc. I didn't think to make it any prettier than that, Day X etc. I could do!
308.
by imdsm 4y ago
I updated my blog post with clarifications but I expect to see two types of changes over time, the variance of output, and the quality of output.
309.
by imdsm 4y ago
Now that is an interesting thought!
310.
by imdsm 4y ago
That is the beauty of the model, that it can understand what a unicorn looks like, and then know roughly where to place the features. Currently we're seeing poor results due to safety tuning but in time, it'll be very interesting
311.
by imdsm 4y ago
The former yes. The expectation is that updates will start to become more and more iterative over time. For this reason, I record the specific version that is pegged to the generic `gpt-4` model.
312.
by imdsm 4y ago
Author here: these images are using `gpt-4` but I'm recording the specific model that OpenAI use with each result. As the incremental updates come out, that will change (without requiring me to change anything.)
313.
by imdsm 4y ago
The model being used is `gpt-4` but each result records the model that was used to generate the output. With incremental updates, the model will change automatically. This is to support incremental updates without needing anything from me.
314.
by imdsm 4y ago
That's a really interesting question! I haven't yet gained access to the enhanced chat features with image outputs. I'm using the API with default parameters, with the gpt-4-0314 model, outputting SVG. It's not a perfect
315.
by imdsm 4y ago
Author here — thanks for explaining this. My expectation is that there will be incremental updates to the model, so while I'm providing the model `gpt-4` for completions, I'm recording the actual model, `gpt-4-0314` in this case,
316.
GPT Unicorn: A Daily Exploration of GPT-4's Image Generation Capabilities(adamkdean.co.uk)
158 points by imdsm 4y ago|106 comments
317.
by imdsm 4y ago
Train people to use GPT-related tech to become superhuman team members. You can do it all yourself, sure, but that doesn't scale. You can either train GPT-tech to use GPT to scale or train humans. Right now, reliability is the most dif
318.
by imdsm 4y ago
I disagree with you. I've used it in production for almost a decade now. There have been issues sure, but name me a single production-level technology that hasn't had problems? - Microsoft Windows: Various versions of Windows have
319.
by imdsm 4y ago
The company has never been good though. The technology was great, but they never found a way to monetise it properly, and their whole approach to outreach, community, and developer experience was terrible. Their support, non-existent. In fa
320.
by imdsm 4y ago
So basically, chlorine compounds (hydrochloric acid) in the atmosphere from CFCs (old aerosols, fridges) turn back into chlorine thanks to the smoke particles, which wouldn't otherwise happen, and the sun then causes the chlorine to br
321.
by imdsm 4y ago
Brexit won't have helped, but I'm sure supermarkets price gouging through inflation is a worse contributor to this. When they're stonewalling farmers with low ball pricing, there's no incentive to produce.
322.
by imdsm 4y ago
Let's not ignore the voices of UK farmers too, who are switching off greenhouses because supermarkets are paying below cost prices, don't negotiate, and don't pass on profits. If supermarkets increase prices by more than 50%,
323.
by imdsm 4y ago
I wonder how many bugs are found this way, rather than actually looking. I found a way to extract private emails from npm a while ago, purely through chance.
324.
by imdsm 4y ago
This becomes the issue. If the school was there to teach you to reach an outcome, and the desired outcomes were difficult enough to achieve, then any (ethical) route you take should be deemed okay. For example, a school teaching you to buil
325.
by imdsm 4y ago
The problem is education, not ChatGPT. Education needs to evolve, and so far, ChatGPT has been fantastic as an educational device. Compare to my experience studying for a year at Open University where the tutors were unresponsive, unhelpful
326.
by imdsm 4y ago
I couldn't say this any better. Once you have kids, everything changes. The impact of getting ill is no longer how it affects you, if you've done enough with your time, but how it will affect them. It's like a dark cloud that
327.
by imdsm 4y ago
Fuck cancer. I wish you the longest possible time.
328.
by imdsm 4y ago
Why is it that some people react so strongly to the concept of therapy? Must be in need of therapy
329.
ChatGPT (Chat.openai.com)(chat.openai.com)
1 points by imdsm 4y ago|0 comments
330.
ChatGPT Free Research Preview(chat.openai.com)
1 points by imdsm 4y ago|0 comments

More ›