vladoh

This looks great, thanks for open-sourcing it!

How does it deal with the current guardrails 5.6 Sol has on finding vulnerabilities? When I use it in the Codex app it would sometimes say it found a vulnerability, but it cannot tell me what it is.

Here is mine built like a VS Code simulator:

https://haltakov.net

I'm working on a way to make it super easy to create and share beautiful photo galleries that tell a story. Take your folders with photos and create a web gallery that works great on all devices.

The project has a CLI interface that is free and open-source, but you have to self-host the gallery. We are also building a SaaS app which is basically a managed version of the open-source tool with a visual builder and we take care of the hosting and CDN.

https://simple.photo

Just a simple tool that I need to optimize images for fast loading by creating different versions in different formats and resolutions

The CSS for this is indeed tricky. I figured out this layout 5 years ago in the v1 and forgot how it works, just took it over as it looks good. The key is that not all rows are exactly the same height. There are small differences that allow photos to fit horizontally.

I also tried the vertical masonry layout, which looks good, but makes no sense if your photos have a chronological order...

The magic happens here: https://github.com/SimplePhotoGallery/core/blob/a3564e30bcb6...

I'm working on a open-source tool to create photo galleries from a folder of photos: https://simple.photo. It creates galleries as static sites that are easy to self-host.

I started this out of frustration that there is no good tool I could use to share photos from my travel and of my kids with friends and family. I wanted to have a beautiful web gallery that works on all devices, where I can add rich descriptions and that I could share with a simple link.

Turned out more people wanted this (got 200+ GitHub stars for the V1) so I recently released the V2 and I'm working on it with another dev. Down the road we plan a SaaS offer for people that don't want to fiddle with the CLI and self-host the gallery.

I think this is the first messenger app that supports this. This is huge for AI chatbots!

We've been using Bluedot for couple of months already and it has been a huge boost in productivity. We often go back to meetings (especially with clients) and review things. We are already starting to use it as a knowledge base for our dev team.

I worked at BMW for many years. I didn‘t do anything relate to their F1 program but once I talked to an engineer that did.

He was a mechanical engineer that designed different parts for the road cars. Once, people from the F1 team went to him because they had an issue with a part breaking frequently and asked him to improve it.

He designed another part and told the F1 guys that this will easily last for thousands of hours. They weren‘t happy. They told him they only wanted the part to last for 2 hours and the one he designed was too heavy. He had to do it again…

One aspect I didn't see much discussion on is about tradeoffs. Yes, computing demand is increasing, but does it all have a negative environmental balance?

Example - modern cars. They are full of electronics (there are around 80 ECUs in a modern car), but these electronics make the cars safer and lower the fuel consumption. Just look at how much the CO2 output for the same engine power decreased over time. Some of it comes from better mechanics like turbos or filters, but a lot of it comes from better electronics in the engine ECU.

Another example - logistics. Delivery companies are able to much better plan demand, routes and inventory thanks to big data processing in data centers. Yes, a data center consumes a lot of energy, but it also helps reduce emissions from trucks driving around less.

I don't have exact numbers, but I think everybody can imagine how increased computing in some areas (not all) improves the environmental impact. These are things that need to be considered in order to avoid unintended consequences.

While driving, do you understand how the engine ECU of your car computes how much fuel to inject into the cylinder, how the ECUs distribute the power, and the braking force on individual wheels, or how your rear wheel steering calculates the turning angle of the rear wheels based on your seed and steering input?

You don't need to know most of the details how your car works in order to drive. You need much more knowledge to build one, yes, but not to drive.

There are different levels of abstractions and depending on your problem you need to understand them only up to a certain level. And different people have different problems to solve.

In most real-world problems today, the difficult part is indeed the data, not the underlying math of the activation function, loss function, or optimizer. Just Google "data-centric AI Andrew Ng" to read more on the topic from one of the most well-known people in ML.

This is a very interesting idea, even if not really practical. The author has more details in the paper: https://arxiv.org/abs/1904.12320

Participants really receive value and I think this is why the model works. However, they don't receive the value they deserve...

I agree that it works, that's why there are so many popular communities. However, how much better it will be if people can actually pay their bills with this work?

I have very fond memories of Delphi - it got me into programing and I earned my first money from a job using Delphi. Nowadays, while doing web development I still miss some features. I haven't used Delphi after version 7 (just before the .NET switch).

However, Delphi is still used by some companies! My mother works in a small company developing business software like accounting, payroll, inventory tracking etc. They still used Delphi (I think version 7) and have a small, but thriving business. No incentive to switch...

Yes, it will be really interesting to read a post-mortem of this.

I also got a notification just saying "foo" 5 minutes before that and thought that was somebody testing in production.

Even if what you argue is true, it doesn't follow from this report. Why is the accident rate of Tesla with Autopilot and all safety features off 2x better than the national average? Because there is a difference in the demographics - Tesla drivers are probably younger and more enthusiastic about driving than the average driver.

Now, if you do the same statistics on the same demographics for all non Tesla cars, you could actually get less accidents than Tesla - here are where the hidden accidents went. Again, I don't have the data about this and I don't claim it is true, but without knowing this, you cannot make the conclusion you are making as well.

Otherwise I agree with you - I also believe that Autopilot+human is safer than just human. Unfortunately, the usual way that people interpret these numbers is that Autopilot is safer than human...

Interesting - I never thought about this aspect! This crash was of course 100% preventable by... driving.

Some people are quoting the recent Tesla safety report [1] as evidence that Autopilot is on average much safer than a human driver. This is a classic case of the Simpson's Paradox [2].

On the first look it seems that Autopilot is 4x safer than driving without any safety features (1 accident every 4.19 million miles vs 0.978 million miles). However, the data used to compute the stats is different in two important ways:

1. Autopilot cannot be always activated. This means that is some particularly difficult situations, the driver needs to drive himself. These are more dangerous situations in general.

2. If a driver disengages Autopilot to avoid an accident and engages it again straight away on a 10 miles drive, then you will have 9.99 miles driven on Autopilot without accident. The statistic misses the cases where the human driver intervened to avoid an accident.

This means that we are comparing the same measure (accidents) on different datasets and therefore in different conditions. This is dangerous, because it may lead us to wrong and often opposite conclusions (see Simpson's Paradox [2]).

I'm not saying that Autopilot isn't safer than a human driver, given that the driver is at the steering wheel and alert, but that this data doesn't lead to that conclusion. If the driver is not sitting at the driver seat, then it is certainly much more dangerous.

[1] https://www.tesla.com/VehicleSafetyReport [2] https://en.wikipedia.org/wiki/Simpson%27s_paradox

We did that in Bulgaria back in the days as well. We also had a special verb for the act of ringing and hanging up immediately - "to clip" (from Calling line identification presentation = CLIP).

There was also the case where one of your friends would have many free minutes in one of the network. Then you would just "clip" them and the expectation was that they call you, because it was for free.

Yes, this will probably happen. Many people in Germany already didn't want it, because of it's worse efficiency...

I find it fascinating that the behavior emerges from the cooperation of many, many people working together towards a common goal.

Yeah, it's a 2 GB file, so it hits the quota limit on the Google Drive pretty fast. There is a problem with the mirror at transfer.army that I'll fix today.

This is one of my projects using OpenAI's CLIP model to do interesting things. So the big credit goes to them!

I processed 2M images from the Unsplash dataset with CLIP and stored the feature vector representation of each photo (a 512 element vector). You can now encode a text query with CLIP in the same latent space and search the database.

I also tried doing some simple arithmetic with the feature vectors to combine the result of a search query and a photo. You can for example search for "Sydney Opera house" and give it a bight photo and you will get a photo of the Sydney Opera at night.

You can directly jump to the Google Colab notebook if you want to give it a try: https://colab.research.google.com/github/haltakov/natural-la...

This is a small side project I built to see how people play the game of Iterated Prisoner's Dilemma. It is funny to see how many people concentrate on beating the bot, instead of maximizing their score...

You can find the bot on Twitter here: https://twitter.com/DilemmaBot. You can start playing by tweeting "@DilemmaBot play".

I'll be happy to answer any questions!

This is a small side project I built to see how people play the game of Iterated Prisoner's Dilemma. It is funny to see how many people concentrate on beating the bot, instead of maximizing their score...

The code of the bot is available on GitHub: https://github.com/haltakov/prisoners-dilemma-bot

I publish the high-score list regularly on Twitter: https://twitter.com/haltakov/status/1364349097554116615

I'll be happy to answer any questions!

I think we will get there soon! CLIP is a new model that OpenAI published in January and I'm sure Google is working on similar technology, which can be used for both video and image search.

You feel like this, because you are right :) Just search online for "google captcha self driving".

CLIP is quite a new model published by OpenAI in January. Their work is novel and pushed the state-of-the-art in this area by a lot. I'm sure that Google is also working on similar applications.

And I think they already have something similar. Recently, I've seen search results on Google that point you to a specific time of a YouTube video...

Yeah, indexing all YouTube videos is surely possible, but out of the scope of a personal project :)