Are Police reports used as evidence in criminal trials? Do we need to add another layer of scrutiny to the process to ensure everything was recorded properly, and not transformed by this model into something that doesn't resemble the truth in the least? We already need to deal with finger print experts and other forensic evidence fraud. Now the police report that's generated will be a source of attack for the defense, along with every other weak investigative tool Police currently use. Get into deep learning now, or 5 years ago and become an expert witness for fast cash.

I use it to generate functions in javascript for testing issues on webpages. One example is that sometimes analytics software can have issues when an iframe from the parent window domain is served. This can lead to doubling up events. I asked chatGPT to generate a function that checks for iframes on the page with the same domain as the parent window. It output something that did exactly what I wanted really quickly. Also these smaller functions are easy to spot check for bugs, so it definitely works for that scenario.

it doesn't help that the redacted information likely contains the largest acronyms known to man that decompose into coded words for a project anyway removing any sort of probabilistic ad lib effectiveness. This is AI generated hopium.

If I can't have Balenciaga Pope, then what's the point of an image generator? I don't want your product if it's just a nerfed image gen in a collard shirt. Go back to innovating on web crawlers and under handed threats of lawsuits to any random domain owner on the web.

I'm surprised, impressed actually, that their lawyers could cobble together a generative AI model.

They have walked back their crazy pricing from years past to something somewhat more palatable. That may be the bigger takeaway for me. Seems like there was some analysis on consumer behavior for the last few models and they pulled back on 1500$ phones.

so pixel tracking is a classic, used by Google Analytics and most analytics packages on the web. It's not "hiding" anywhere, transparency doesn't matter, you don't even return any data for the "pixel" the 1x1 gif request just gets a 200 back. The get params and body of the request have all the info that are going to be stored by the server getting the pixel request. Part of the reason this is the case is that it used to be that gif requests had the shortest response times on the web and skipped some steps.

Small plug here. Ezoic is working on providing a tool for self hosting videos with their own central domain. When a user clicks on a video in the search section of https://www.humix.com/ it will then point the user to the domain of the video content owner. Providing a way for content creators to have ownership over their work and ad revenue while still being on a central, searchable and social network site. https://www.ezoic.com/humix-ne...

We want to give content creators and publishers control back over their content, moderation and ad revenue.

This is a terrible take. Golang is used in lots of large production environments. Twitch, discord and docker to name just a few. This is not some toy language flavor of the month bullshit like you claim. It is a better option than JS for all web server applications. Package management can be done using actual GIT HASHES instead of arbitrary version numbers through npm/yarn. It is portable and doesn't run on a VM. It is SIMPLE and forces proper error handling. Honestly, this sounds like a post from someone who's never worked with go and has only viewed it from the outside in.

Google, Facebook, probably amazon and twitter will all soon need new ways to preserve ad performance and targeting to maintain their rates. The "solution" seems to be the browser. There will probably be more effort in this space as litigation becomes more of a concern. Current browsers are mostly owned by companies that seem to have a vested interest in this kind of tracking though. Brave seems to be the "best" option even though they don't seem to be very well run, will multiple weird miss steps and a crypto currency tie in. Google has the best position since they have chromium which both Microsoft and brave base their browsers on. There's Opera but that was taken over by some weird company and is now total shit.

I hope Meta doesn't acquire mozilla and create a new type of browser war.

This kind of physical machine learning model was something I always wanted to try making for a local science museum. The idea was to create a 32x32 grid of boxes and allow kids to fill the boxes with small BBs in the general shape of a letter of number, then let the BBs get siphoned through a series of physical gates whose weights were adjusted based on the MNIST data set. The bottom of the machine would be 36 buckets representing each letter and single digit. The BBs would then pool into the buckets and hopefully show the letter the child wrote as the bucket with the largest number of BBs.

It'd be interesting to see if they could create a kaleidoscope type object that could be pointed at an image or object and do simple object recognition with just the incoming light.

web person here, mostly use golang. Java can kick rocks.

As someone who has unloaded trucks for Target during over night stocking, I can say I would have loved one of these. It's never a certainty that the people who loaded the truck properly loaded it from bottom to top, heavy to light. So there was always the possibility something was going to fuck up your night and the rest of your shift. Less worker injuries is also awesome. They don't get the best and brightest to unload these trucks, so poor lifting habits got to people over time. Good stuff.

Interviewing 100 candidates to get 1 person really isn't that many hours to burn, especially when you want to make sure you're hiring people that can function independently and do not need to be babied and watched and code reviewed at every step. Having a process that asks algorithm questions can lead to a higher number of false positive hires. We're trying to avoid that. 100 applicants - resume filters out 30-40 candidates (with fun cases like resumes geared towards art or biological sciences getting in the mix), then phone interviews filtering out 20 candidates, technical interviews run by a third party ByteBoard and maybe 5 finals interviews from that first 100. I wonder what you think the success rate is for many other startups/companies?

Except I'm not seeing older individuals ace our technical rounds at all. We have hired two 50+ engineers over the last two years (of 40 hired) and engineers in that age range applying are very low to begin with. We are a startup that is growing which may reduce the likelihood of an older individual applying (why get into a startup late with low options when you can apply somewhere else?). I just haven't seen a bunch of older individuals crushing our technical assessments, again they language and technologies they use are up to them so it's not a lack of knowledge of "modern" languages or front end html frameworks (vue, react, angular).

We have two senior software engineers proctor our final technical interview process, which involves asking the candidate to describe the code they had written and then make changes that we request at the time of the interview. That being said the success rate for anyone applying is at or below 1% . Doesn't matter the education level either. Bachelors, Masters or PHD the success rate is similar.