nl

24 ms·

571.
by nl 7mo ago
NYTimes had a story about this (March 12): > Meta’s new foundational A.I. model, which the company has been working on for months, has fallen short of the performance of leading A.I. models from rivals like Google, OpenAI and Anthropic o
572.
by nl 7mo ago
This seems pretty good.
573.
by nl 7mo ago
> They were ~gpt4o, with the added benefit that you could run them on premise. No, they are bad models. They were benchmaxxed on LMAreana and a few other benchmarks but as soon as you try them yourself they fall to pieces. I have my own
574.
by nl 7mo ago
The maths shows inference is very profitable. Look at how Google/AWS/Azure change the same rates as Anthropic does for running Claude models.
575.
by nl 7mo ago
Anthropic has raised $64B in total since they were founded. Even if you say we are going to measure profit in the very special hacker news way of looking at money taken in from customer revenue against money invested and we say they c
576.
by nl 7mo ago
Well Opus 4.5/4.6 kinda was right? I mean software development has changed more since then than it has in my 30 year software development career.
577.
by nl 7mo ago
> And to be clear, OpenAI/Anthropic most definitely know this: that's why they've been aquihiring like crazy, trying to find that one team that will make the thing. Anthropic is up to $30B annual recurring revenue. I wish
578.
by nl 7mo ago
It's higher than all other models except vs Gemini 3.1 Pro on MMMLU
579.
by nl 7mo ago
> barely competitive It's higher than all other models except vs Gemini 3.1 Pro on MMMLU MMMLU is generally thought to be maxed out - as it it might not be possible to score higher than those scores. > Overall, they estimated tha
580.
by nl 7mo ago
This is just a grab bag of different critisms to what you said. Vendor financing is very standard across multiple industries. It's not a loan in the sense you originally stated because it's physical assets, not financing. The circ
581.
by nl 7mo ago
Doesn't the framework desktop have a Ryzen 395 AI? That's a unified memory architecture like the Macs.
582.
by nl 7mo ago
This isn't right > It's largely SoftBank, Oracle, Microsoft and Nvidia, all of whom don't have big piggybanks full of hundreds of billions. Actually SoftBank, Microsoft and Nvidia literally have free cash sitting there. N
583.
by nl 7mo ago
It's a valuation, not the amount they raised.
584.
by nl 7mo ago
> Plus I think most people using agents for coding are using subscriptions which they are definitely not profitable in. Where on earth do people get this idea? Subscriptions that are based around obscure, vendor defined "credits&quo
585.
by nl 7mo ago
The medium one on OpenRouter didn't support tools when I tried it. I will update when there is one.
586.
by nl 7mo ago
Yes exactly. I doubt they actually want to do this. They clearly see having a wide set of paying customers as valuable (otherwise they'd just raise prices) but if you are stuck having to make hard choice then I can see the attraction
587.
by nl 7mo ago
I suspect people are misdiagnosing the root cause of why Anthropic is doing this a bit. I don't think this is particularly about the financial impact of people using OpenClaw - they can adjust the amount of tokens in a subscription qui
588.
by nl 7mo ago
I doubt anyone actually thinks the Iranian regime is good in anyway. But I thought the whole points of MAGA was "No new wars". And now there's a new war, without any real reason (other than something something Netanyahu and t
589.
by nl 7mo ago
Xiaomi's mimo-v2-flash is great if you care about speed and performance - it's 1/10 the price of Gemini 3.1 Flash Lite and faster (on OpenRouter). GCP does server other non-Google models, but I'm not sure what they have
590.
by nl 7mo ago
23/25 on my agentic benchmark for the free version on OpenRouter. That's a great score - only 4 models have ever scored higher. But there are open models that also score 23/25 including Qwen 3.5 27B.
591.
by nl 7mo ago
Gemma-4-E4B-it scored 15/25 on my https://sql-benchmark.nicklothian.com/#all-data (agentic SQL generation). The naming is a bit odd - E4B is "4.5B effective, 8B with embeddings", so despite the name it is pro
592.
by nl 7mo ago
Concentrating on LMAreana cost Meta many hundreds of billions of dollar and lots of people their jobs with the Lllama4 disaster.
593.
by nl 7mo ago
Nemotron-Cascade-2-30B-A3B is worth checking out too at the size - I found it even better than Qwen 3.5 35B! I ran it slowly but successfully on a 8GB 1070GTX with CPU offload.
594.
by nl 7mo ago
Listen to the podcast I linked. 2 years is ramp up time for new memory factories (which are currently being built)
595.
by nl 7mo ago
The OG Lllama 3? GPT-OSS? NVidia Nemotron 3?
596.
by nl 7mo ago
Trinity Large Preview managed 17/25 on my agentic SQL benchmark: https://sql-benchmark.nicklothian.com/?#all-data which is a fairly mediocre score for a large model (Qwen 27B managed 23/25) This non-Preview releas
597.
by nl 7mo ago
Mimi Flash matched Mimo Pro on https://sql-benchmark.nicklothian.com/?#all-data at double the speed and for $0.003 instead of $0.07
598.
by nl 7mo ago
I had an iPhone 12 and just upgraded to a 17 because I don't see any point waiting - prices will just increase. > It really sucks more because the reason for it--AI--is just so godawful and pointless. Strong disagree. AI is the bes
599.
by nl 7mo ago
In the Dwarkesh podcast with Semi-Analysis's Dylan Patel they forecast the phone market will shrink by 50% this year because of RAM prices: But that’s the high end of the market, which is only a few hundred million phones a year. Appl
600.
by nl 7mo ago
> US tech companies just continue operating because "revenue and growth". US tech companies are some of the most profitable business in history. Google made over $130B profit last year, Meta 60B. I'm old enough to have had

More ›