mrinterweb

We all know who Trump is. This behavior, from Trump, should shock no one by now. I'm equally upset with all those who enable him. Midterms are here, time to do the one thing we are still allowed to. Be certain, Trump will cry foul on the elections unless they go his way.

There's no way I would be able to detect if this was AI from the examples alone. If I got a phone call from a malicious agent using this tech, I probably wouldn't realize it. My plan is to ask a challenge question about a shared memory with people I personally know, if there's something suspicious about a call. I don't know how else to prevent it. It sucks to be this suspicious, but I can't help it.

There's so many potential scenarios. One scenario: An AI voice trainer trains on someone's social media video posts and get a pretty perfect copy. Then SIM jacks the target's phone number and calls pretending to need help. AIs could be used to mock ransom people, request bail money, ask for a change on a production server, etc.

People with high public visibility are likely targets, especially if they have money and or power. Impersonating them to subordinates would be an easy means of a social engineering attack.

I know the doomerism of robots takin R jobs get repetitive, but this seems like it would work very well in call centers. I get frustrated talking to AI call center agents, but that's often because they are less capable and prone to bad logic errors. To be fair, speaking to humans in call centers can be annoying too. If a call could be answered without waiting in a queue for 30 minutes by a competent AI agent, I think I would prefer that.

That probably explains why there were so many competitors around withing days of the Jev announcement. They are not starting with a moat, and there doesn't seem to be any moat in sight. Just buzzword recognition because everything is comparing to "jev".

Curious, what hosted VPS services people are using these days. I am using hostinger, but I don't intend to continue as 63%-70% discount eventually runs out. Then the real cost hits. Its the lock-in and migration costs that make people shell out the uncomfortable renewal price. Looking for an affordable option without the introductory rug-pull.

Projects like DwarfStar https://github.com/antirez/ds4 really lower the hardware bar a lot so Deepseek 4.1 flash and other mixture of expert models can run on consumer hardware. There are also other inference providers who make their money serving openweight models. Services like OpenRouter make it all too easy to utilize these models. Access to these models isn't hard. The hardware moat is becoming pretty easy to bridge.

I'm giving this a go, and so far, this is great. I have 1x RTX 4090 (24GB VRAM) and 128GB DDR4 RAM. I am seeing > 110 tokens/sec (3 token MTP). Using 60K/260k context currently. So far the results seem at least on par with my Qwen 3.8 q5 27B (~60 TPS). I realize ultimately tokens/sec don't really mean much if the quality sucks, but I am optimistic but still paying close attention to the results.

Working with fast local models can be great. Fast prefill, and >100 TPS is quite quick.

I assume the GAMERFAN is applied to ceiling with thermal paste

And yet people will let AI agents run autonomously on their machines. I feel like we're reaching peak YOLO with security.

Rails needs hard break from DHH. He has been toxic for the rails community for a long time. I sincerely appreciate his many contributions to rails, and his shaping of rails, but it is beyond time to cut ties. He's clearly not interested in rails any longer.

DHH founded Rails World (as far as I know). What a wonderful way to tell the whole community to F'off than to sell tickets to your conference then tell them not to use the technology. If DHH wants to vibe code everything in rust/assembly or whatever claude tells him to do, let DHH do his own thing, and let's part ways.

If inference continues toward the trend of becoming a commodity, and inference hardware efficiency is doubling every two years, that will likely mean that the price of inference will continue to fall so long as the market is competitive. I don't see how companies investing billions in inference hardware today will see profit in the lifespan of the hardware.

One aspect I am bothered by is how all writing style is becoming homogenized. Individual writing "voice" is becoming one collective AI voice.

I hope Apple contributes to an open standard for this, instead of keeping this proprietary. In the emerging world of generative AI, we need this more than ever. Not just iPhone, but ideally most camera systems will have some kind of verifiable capture mode.

The timing of the AI doom narrative could not come as a more suspect time. Chinese labs are releasing open models that are nipping at the heals of the frontier models. Anthropic and OpenAI know these frontier class open weight models can threaten their valuations by changing inference to a commodity where the provider who can serve the best models the cheapest and fastest will win. Few are going to pay $50/million tokens when they could pay <$1/million tokens for similar outcomes. There's little hope investors could recoup the wild investments made into Anthropic and OpenAI, IMO.

The benchmark comparison to other models was suspiciously missing. The self-comparison is a good representation of progress, but it is light years behind frontier models. Speed is great, but wrong is much worse than slow, IMO.

> triggered memories

Yeah of 10 minutes ago. It is shocking how long some seemingly simple things can take. I know there are some things I can do faster than the LLM and some things it can do faster than me. The amount of rambling BS is the exhausting part.

I think lithium air batteries will be the real inflection point for aeronautics. Li-air has potential density 12 kWh/kg. CATL is focused on this tech. May be some years before they are able to achieve the full potential density. Still, as the other commenters said, the efficiency of electricity make make up that difference.

I saw the version of this video with Paul Rudd (Celery Man) https://youtu.be/a8K6QUPmv8Q?si=TWmoNhxYAPp73TKg

I'm really curious how GML-5.3-flash would do. Very affordable, and it seems to do pretty well with 3D modeling.

That's wonderful. I was going off an older version of the Artificial Analysis page for GLM-5.3-Flash https://artificialanalysis.ai/models/glm-5-3-flash. The page is updated now to show that it does support multi-modal image inputs.

I really wish GLM models had vision capabilities. I've worked around that in the past to use a vision MCP in my harness that GLM can call. It is not the same, but it allows the model to query images.

Give it a couple days, and there will be plenty of other inference companies hosting it. Don't like z.ai's TOS? Use the model on a provider with TOS that you agree with.

The pain points in the article do not bother me. I'm bothered by Opus 5's verbosity. It is so long-winded and you have to read through verbose outputs to mentally distill what is important. It is exhausting. I don't think I've ever started skim reading LLM output more that I do with Opus 5. I use the caveman skill, and I think that does help some.

Exactly. There could be a lot of value for inference companies to do this. Could save a lot of money being able to hand off highly repetitive known tasks to far smaller specialized models.

There is so much opportunity for purpose built models like this. Ideally a harness should spin up a subagent to offload to targeted models for specific tasks like this. I know this is not a novel idea. Claude code does some of this by handing off the "explore" agent work to haiku. I just love seeing that specialized LLMs are being developed.

This looks fantastic for a common async workflow I use. I often use one job to fan out multiple individual http request jobs. The reason I prefer jobs for this is easy and consistent retry logic, and durability. I want to make sure those HTTP requests eventually go through. Fibers would be much better suited for this. So much of work that goes onto work queues is IO bound, and fibers are a great fit for that.

That ruby fiber vs go goroutine benchmark is interesting. The 4-10x memory use doesn't surprise me, but the near performance does. I'm guessing there is more of a gap with the p50.

Look at robotics demonstrations on YouTube from one or two years ago, and compare that to where we are today. It seems reasonable to me that general purpose robotics may be able to outmaneuver your average human in a few years. At least one robot has surpassed what I could do https://www.youtube.com/watch?v=Xha2TJ_1C6Q

Two that I use are:

* Openrouter.ai for a hosted router

* https://github.com/diegosouzapw/OmniRoute for a local router

The don't only host open weight models. Also, why not promote this. If Fireworks thinks this big news might convert some new business doesn't make it not true.