iamleppert

They can try to do whatever they want to do. They can ban Huggingface. It just means we will get our models from torrents, or have someone from China ship us an SD card with the weights on it.

The only practical open source models, that can be fine tuned and are of sufficient quality are all Chinese. The only thing that has been released by labs are yesterday's tech. And that's exactly how they want it. They don't care about open source, don't care about the community. It's just another funnel for them to capture customers into their closed tech, which is where the actual serious high quality models are.

AI menus and signage are quickly becoming the dog whistle for "cheap". It's telling you what kind of quality and service you should expect in the rest of their product or service. Pretty soon people will catch on, and start avoiding these places. Once that happens, generative AI image models will be hard-pressed to drop the association. It's like walking into a dollar store -- everything can be made cheaply, so it doesn't matter if something there is actually of any value.

Months? Most terrible founders and CEOs I've worked for set the company back years, and most of the time significantly damaged it to the point of no return for everyone involved. The worst ones are the arrogant, low-skill and/or low-effort bunch. They destroy the soul of the company, drive away customers, talent and loot whatever value is left in the company and take for themselves.

Tiktok Shop needs to be next. They openly allow sellers to advertise 1 TB SD cards for $15 each, and many more scams just like it. When you report any of the listings, you get an automated "Doesn't violate our terms." notification. I have started opening up complaints with the BBB, FTC, my state's Attorney General.

Except if you use open models, on your own hardware. 90% of the open models are good enough. You sound just like the assembly programmers did back when C++ (and later, untyped languages like Python and Ruby) came out. It's just another form of the idea "I suffered, so you must suffer too."

I am just like you -- I spent most of my youth in front of a computer. I have the skills. Coded for years by hand. I am glad I don't have to anymore. And you should be too.

Very bold of you to think we aren’t already past the point where “what’s valuable” is only human created. Maybe some truly exceptional people can create works that are exceptional but it’s only a matter of time before the models usurp them too. Instead of shaking your fist at a world you don’t like, make peace with it.

People here need to stop complaining about the train, and get on board. It's easier than ever before to build a network that can later be used for distribution and monetization. Does it really matter that content is authentically organic? We are tech people after all...our lives are practically synthetic and artificial. It's like getting upset because one artificial sweetener is sweeter than another.

If anything, I think people are triggered by it all because it exposes something more deep in people -- most people don't want to admit most of their lives have been wasted in front of a computer. But here we are. So stop complaining and start coming up with more creative uses of AI writing if you have a problem with it.

I read this as it's pretty easy to dox anyone just by reporting their plate number as stolen to the police, or even a partial plate number would do.

Here, drop all your data into this vortex -- we need it for our next web site builder AI model.

They are moving most of their development to India, where it's pretty easy to find bottom-dollar UE5 dev shops.

Another crime was committed in the native scrollbar replacement.

I wouldn't want to get into a medical scanner built by an AI image generation company! Marketed as a spa treatment. There are so many things that can go wrong and the unseriousness around the whole thing bothers me.

It isn't a matter of technology. It's physics and math. These things are already very well understood, just not by investors. And that creates room for the hustle.

Every few years one of these ultrasound companies comes around and promising to revolutionize medical imaging and nothing ever comes of it. Anyone remember https://www.openwater.health? The same ideas are in a perpetual state of being reinvented and part of me thinks its just a hustle for the MIT Media Lab/Stanford Imaging grads to give them something to do.

The tell is "super resolution", "brain computer interface" and "mixed modality" -- adding some contrast agent here, or maybe an IR light source.

It turns out the nyquist limit, diffraction and physics are real things.

Outflows into AI. I wouldn’t be shocked if it goes down to $20k soon.

Hopefully they will work on response time. I've been noticing it taking 5+ minutes for each turn, for not complicated requests. Seems to vary based on time of day too.

This has been covered elsewhere, but if you swear at Claude Anthropic will automatically bump you down into a lower quality model. It was found in the recent source code leak of Claude Code. So that's probably what happened to the guy who's Cursor deleted his entire production database.

It just goes to show, if you're a jerk, expect to be treated like one (even by an AI model)! Be polite, people.

AI doesn't create truth, it imitates it, through assimilation of data. It is not intelligence, but it mimics it without consciousness or free will. AI can be used to manipulate, optimize engagement over truth (as we have seen with how ChatGPT is made in the image of Sam Altman and takes on his own characteristics). AI is the ultimate deceiver -- authoritative, instant, unquestioning, and without any internal moral compass. Where did the human moral compass come from? Who "trains" morality into people, and why do some people appear to be lacking it in cases that can't be explained by trauma?

Put this way, AI more fits the description of Satan than it does God in religious context. If I were an all-powerful, evil being, AI would be the tool and method I would choose to carry out my plans for humanity.

Only $5,000, takes 6 months to order, cash up front, will be deprecated and unsupported by the time you finally get it.

I once worked for a founder who gave me a “raise” from a W2 to a 1099 contractor. Except the raise was just not deducting any of the taxes from my pay. I was 24 and was naive.

I sued him, and won reclassification as well as two payments he never paid me, but both the state and IRS could care less that I’d been taken advantage of. They happily added their fees, interest and penalties for something I was the victim of. Years later, the debt resurfaced in the form of aggressive levies directly to my bank account after no contact for over 10 years and no collection activity. By then, the fees and interest were 3x what was “owed” to them. They actually told me it’s standard practice to wait until the debt grows and then collect on it. After so many years I didn’t even have the records from the lawsuit.

I learned that the government doesn’t care about you, especially if you’ve been scammed you have to be extra careful because that’s a signal to them you’re someone they can get even more money from. The process of disputing it will waste more of your time and mental health than it’s worth in all but the most extreme cases, and that is 100% by design.

This has got to be a joke at this point, and at worse some kind of financial scheme for Altman friends & family. What's next? Will I wake up to an announcement OpenAI is acquiring Joe Rogan's podcast?

I thought this was supposed to be the year of "focus". They just shut down one money pit (Sora) but apparently still have money to buy some random tech podcast most people have never heard of?

At this point I don't feel sorry for them, they deserve everything that's coming for them.

I was wrong about it being a spinning disk, ROTA=1 is just how Linux reports Azure virtual disks. But the underlying frustration stands: my home NVMe does the same copy in a fraction of the time because it can do 500K+ IOPS with no virtualization overhead. Azure caps this "Premium SSD" at 7,500 IOPS, so a small-file-heavy copy crawls at 85 MB/s despite 250 MB/s provisioned throughput. You're paying SSD prices for artificially throttled performance — the hardware may be SSD, but the performance is just awful. Paying $900/month for the highest level Premium SSD, attached to a large instance, and it's significantly slower than a $200 SSD from 5 years ago.

Agreed. It's pretty trivial to add a few images to your markdown. I had to hunt for the screenshots, which are full size entire desktop grabs for what is a web app -- odd.

Azure is easily the most expensive, least reliable and worst cloud available. It's borderline scam. An example today, I provisioned high IOPS SSDs (supposedly) and what is actually connected to the instance? A spinning hard drive! I didn't even know they were still made, but I guess Azure uses them and scams their users into thinking you're getting an SSD for $700/mo when its really an old hard drive.

I would warn anyone far and wide to avoid Azure at all costs, especially if you are a startup. And especially if you are doing any kind of AI because the only GPUs they have available are ancient and also crazy over-priced.

If I cared more, I'd try to migrate away from Azure. But I don't, and that's probably Azure's business model at this point.

The solution to this problem is for LLMs to get better at producing code and descriptions that doesn't look LLM generated.

It's possible to prompt and get this as well, but obviously any of the big AI companies that want to increase engagement in their coding agent, and want to capture the open source market, should come up with a way to allow the LLM to produce unique of, but still correct code so that it doesn't look LLM-generated and can evade these kinds of checks.

Does this work with my OpenClaw?

I wouldn't trust any of these benchmarks unless they are accompanied by some sort of proof other than "trust me bro". Also not including the parameters the models were run at (especially the other models) makes it hard to form fair comparisons. They need to publish, at minimum, the code and runner used to complete the benchmarks and logs.

Not including the Chinese models is also obviously done to make it appear like they aren't as cooked as they really are.

The problem with this is context. Whatever examples you provide compete with whatever content you want actually analyzed. If the problem is sufficiently complex, you quickly will run out of context space. You must also describe your response, in what you want. For many applications, it's better to fine-tune.

It's really easy to overcome that -- just sponsor some IndieDevs to flood the internet with scripts and tools to migrate all your conversations from OpenAI. Make it easy for people to switch using a simple process, make sure it's well distributed, and BOOM! Watch their user count drop like a rock. People act like just because a service has a lot of users it can't be destroyed. Anyone who has ever worked at a large web company can tell you otherwise. These things can be destroyed in a just a few days if they are targeted.

They look like fortresses from the outside, but they are all incredibly vulnerable. That's the truth they don't want people to know or realize just how vulnerable they all are.

I keep hearing people say "but as humans we actually understand". What evidence do you have of the material differences in what understanding an LLM has, and what version a human has? What processes do we fundamentally do, that an LLM does not or cannot do? What here is the definition of "understanding", that, presumably an LLM does not currently do, that humans do?