rockinghigh

7 ms·

1.
by rockinghigh 1mo ago
Using tools to fix your writing style or grammar is not exactly cheating.
2.
by rockinghigh 2mo ago
For small models, tokenization can reach 1-10% of total inference time.
3.
by rockinghigh 2mo ago
A coding agent driven by a large LLM can delegate smaller tasks to a faster model. For example searching through the codebase for references, examples, or established patterns. They are treated as tools and don't pollute the main agent
4.
by rockinghigh 2mo ago
Open-weight models were lagging 4 months behind OpenAI/Anthropic at the beginning of the year. They are now just 4-6 weeks behind.
5.
by rockinghigh 3mo ago
The US is also falling behind Chinese manufacturing. They had to ban Chinese cars because legacy American automakers couldn't compete.
6.
by rockinghigh 3mo ago
Same here, I fine tune LLMs for specific use cases. How can I trust Anthropic models not to introduce bugs to preserve their moat?
7.
by rockinghigh 3mo ago
Imagine if your IDE started injecting bugs into your project just because your code looked like it implemented a competing IDE.
8.
by rockinghigh 4mo ago
Consumers don't control zoning laws or risk mitigation details.
9.
by rockinghigh 4mo ago
Gemini tends to be faster and the Flash and Flash Lite models outperform ChatGPT's equivalent models.
10.
by rockinghigh 4mo ago
It may be a push from advertisers who want access to this format. Google Search competes for their money against the competition.
11.
by rockinghigh 4mo ago
They own tons of engagement data associated with that index from being the default search engine on most devices.
12.
by rockinghigh 4mo ago
Why wouldn't getting more customers the plan? Anthropic doesn't acquire companies to have a lower market share. There is clearly a consolidation and a rush to get as much of the developer market as possible.
13.
by rockinghigh 5mo ago
The MoE experts are quantized to int4, all other weights like the shared expert weights are excluded from quantization and use bf16.
14.
by rockinghigh 5mo ago
They could release data to back up that claim.
15.
by rockinghigh 5mo ago
Are there any protections from industrial espionage when using Anthropic, Cursor, Gemini, or OpenAI?
16.
by rockinghigh 6mo ago
Their revenue was $57.4 billion last year. Just in Q4; cloud revenue $6.7 billion, cloud infrastructure $3.0 billion, cloud application $3.7 billion, Fusion Cloud ERP $1.0 billion, NetSuite cloud ERP $1.0 billion.
17.
by rockinghigh 6mo ago
It's the number of attempts at answering the question.
18.
by rockinghigh 6mo ago
He founded the team that worked on fasttext, llama and other similarly impactful projects.
19.
by rockinghigh 6mo ago
He founded FAIR and the team in Paris that ultimately worked on the early Llama versions.
20.
by rockinghigh 6mo ago
It can also be used to simplify existing code bases.
21.
by rockinghigh 7mo ago
It's a lot simpler. These models are not optimized for ambiguous riddles.
22.
by rockinghigh 7mo ago
How is this riddle relevant to a coding model?
23.
by rockinghigh 7mo ago
They do. Pretty much all agentic models call linting, compiling and testing tools as part of their flow.
24.
by rockinghigh 7mo ago
It's called problem decomposition and agentic coding systems do some of this by themselves now: generate a plan, break the tasks into subgoals, implement first subgoal, test if it works, continue.
25.
by rockinghigh 7mo ago
A language model in computer science is a model that predicts the probability of a sentence or a word given a sentence. This definition predates LLMs.
26.
by rockinghigh 7mo ago
How do you join two datasets using r-trees? In a business setting, having a static and constant projection is critical. As long as you agree on zoom level, joining two datasets with S2 and H3 is really easy.
27.
by rockinghigh 7mo ago
I wouldn't say R-trees solve the problem better. Joining multiple spatial dataset indexed with r-trees is more complex as the nodes are dynamic and data dependent. Neighborhood search is also more complicated because parent nodes overl
28.
by rockinghigh 7mo ago
That's not true when tiling the Earth though. You need 12 pentagons to close the shape on every zoom level, you can't tile the Earth with just hexagons. That's also why footballs stitch together pentagons and hexagons.
29.
by rockinghigh 7mo ago
Instagram uses it as their main backend. They have hundreds of million of daily users. Some of the critical backend services are in C++.
30.
by rockinghigh 8mo ago
Tariffs were a huge point of debate in his first administration. The government had to pay $30 billion to farmers to offset the impact of tariffs. > China implemented retaliatory tariffs equivalent to the $34 billion tariff imposed on it

More ›