__natty__

With this at 1k tps and Kimi 2.6 1k tps by Cerebras, I believe we are entering the next stage of LLMs, where companies will also compete on throughput

Lmao. Price for the original battery pack is absurdly high. Good clone.

Maybe I’m naive but the longest single workflow I ran was maybe 15 minutes. How do you steer agents to run “overnight”? And what is the quality of such execution?

It’s fascinating for me to see how small language models grow recently in capabilities while still consumer friendly in size to run on their machines

It's good there is a new player on the market, I take benchmark tables with a grain of salt, however. Speaking about model presentation it's funny to see how clearly their website is inspired by other AI company blogs with extra innovation of hijacked scrollbar.

Maybe one day companies will optimize AI costs by hiring people?

> Together, we are organising around the things we want to change. Starting with: Pay transparency Flexible working An end to crunch

That’s a lot of demands, what next? Competitive salary?! /sarcasm

I hope more people will start fighting together for better work conditions. Company owners have money and lawyers so workers must unite to fight them back. I’m saying this as employer.

As for the demo it's fast and extremely dumb like expected for 2B. I asked how to stop drinking habit and in just one follow-up message it recommended trying 8% ABV. Hilarious.

Why not for 3 eur buy some basic arduino or other tiny hardware to tinker with and for another few eur, tiny i2c/oled display, wires and set of basic switches? You start programming with option to expand to the larger project in the future. You have constraints of real device, community is much larger and there are more learning resources.

So it makes sense to first send stable prompt, reasoning and files content, tool calls summary and actual tool calls at the very end?

I wonder when we reach speed of 1000 tps with high quality models. 5 years? 10 years?

Sounds different because it’s hidden advertisement not a regular blog post

Deliberate thinking will be a think soon in our field.

This whole situation of forcing people to spend tokens is hilarious and shows who is true tech manager and who is impostor in the field. I think it’s a good indicator of competence and hopefully board members see that soon.

I agree. We focus on content where usually it’s so poorly structured and displayed it lacks final touches. I prefer wider scrolling bar than extra white space in the browser. Same goes for finder. I find old systems to be much more usable for me than modern ones

To add to the argument about European funding, Polish people are also very hard-working and probably have the mentality closest among European countries to what the USA had in the twentieth century with the pursuit of the American Dream. The only difference is motivation. Polish people suffered a lot in the past, so they do not want to experience poverty again; thus, their drive is powered by insecurity compared to an optimistic confidence that hard work would lead to prosperity in the future (this is also seen in the Polish sense of humour, which is much darker). I suppose it is mostly because of the post-communist Balcerowicz Plan transformation and the first generations travelling to the West for work, which further solidified the belief in upward mobility from the lower class to the middle class to the upper class through hard work.

For adults, sure. For kids saving for the console it may be problematic.

Contrast between official [0] and third party status pages [1] is huge. How their terms of service for SLA are legal if they are so different from real world usage of their product? I really like GitHub and their services but every time when it’s broken and their status page is green something screams inside me.

[0] https://www.githubstatus.com/ [1] https://mrshu.github.io/github-statuses/

This. We don’t need AI searching for critical bugs in the software, enshittified life quality is already enough.

Surely it is marketing. It’s some “we are danger” narrative, from Anthropic Mythos and now OpenAI too.

> But interview after interview, the story started to make sense. They were not wrong, I was not ready, and it took me a long time to admit that.

I believe this is one of the most humbling but also maturing moments in career and adulthood.

> Nothing in the contract promises capability or performance.

Taken to its conclusion, Anthropic could silently replace Opus with Haiku quality internals and you'd have no recourse. If that sounds absurd, that's exactly where the legal argument lives. Mandatory consumer protection provisions like on misleading omissions cannot be waived by clicking "I agree." Withholding material information about a product you're paying a premium for isn't covered by T&Cs. It's the specific thing those laws were written to address.

You compare tiny modal for local inference vs propertiary, expensive frontier model. It would be more fair to compare against similar priced model or tiny frontier models like haiku, flash or gpt nano.

New model - that explains why for the past week/two weeks I had this feeling of 4.6 being much less "intelligent". I hope this is only some kind of paranoia and we (and investors) are not being played by the big corp. /s

It's terrible that giant cloud providers such as Google or AWS doesn't allow for hard cap at project levels or prepaid. And that especially because alerts are delayed as author stated "We had a budget alert (€80) and a cost anomaly alert, both of which triggered with a delay of a few hours. By the time we reacted, costs were already around €28,000.".

I like the "european technology" movement not because of any nationalist ideas, but because it stimulates technological innovation and creates a new dynamic.

Not exactly for software (although there is such section) but I use end of life [0] website. Besides time when certain software will be outdated it also tells you their release time.

[0] https://endoflife.date/

I wonder whether this kind of release of model could become the spark that ignites a new digital "cold war" between us, europe, india and china, in which they will try to outwit their rivals and compromise their critical infrastructure using artificial intelligence.

Also I’d like to believe that this really is such a huge step forward compared to Opus, but lately I’ve found it hard to believe when I look at the statements made by the CEOs of AI companies and their associates, who are fuelling the hype surrounding this topic even further. Of course, it is good that large companies and industries that are crucial to the country are the first to have access to this, but until the launch takes place, I will approach this with a degree of scepticism.

I love bunny.net. For my use case it provides lower latency than Cloudflare.

I would love to see Java inspired language compiled to Go. I really like Go portability and standard library and Java... verbosity. I prefer explicit names, types and all the syntax around that. Graalvm is not an answer for me because as far as I'm aware it doesn't support cross-compile.