adityashankar

82 karmajoined 5 лет назад

goto https://river.berlin

comments

вчераdiscuss

I was worried about some messup with taxes (my tax advisor messed up here), I managed to get it sorted on time - this was super naive but in -germany when they send you the taxes they mention the "cents" place as well in a very weird way, I assumed that was the entire number and assumed the tax issue I had was 10x larger than the amount it really was (10x and not 100x as my brain is in a place of pressure due to other circumstances and I wasn't thinking clearly)

GPT 5.6 incorrectly stated that I had nothing to do, Fable got the issue correctly and I was able to see that that was indeed the cents place and that I was more worried than I realized, and I managed to get a temporary solution setup (that I verified and I am sure is correct).

Which is to say it managed to relieve me of quite a bit of stress haha

позавчераdiscuss

How do you even estimate that, human productivity is non linear to worktime and often, employees, when given good benefits are likely to be more productive.

also there's the fact of this having a ROI since people generate economic value in the longer run, and hence, more taxes

8 дней назадdiscuss

Kramnik has a habit of accusing people of cheating without evidence https://en.wikipedia.org/wiki/Vladimir_Kramnik#Anti-cheating..., he has also done this to many many individuals without evidence (and cheating is taken very seriously), this is in connection to that - he's being penalized for accusing people (repeatedly) without evidence.

Just to be clear – this is an event where he has been doing so repeatedly, accusing people of cheating without evidence in-and-of itself isn't forbidden - since cheating in chess is difficult to verify, it's the manner and fashion to which Kramnik has done this that has led to his suspension

10 дней назадdiscuss

Location: Berlin, Germany

Remote: Everything works, I'm open to remote, in-person and hybrid working opportunities

Willing to relocate: No

Technologies :

Pytorch/Tensorflow/Vertex AI/Amazon Bedrock/hf-related libraries, primarily ML-based, but otherwise I'm also familiar with Docker, Python, Pypi, uv, Nodejs, PSQL, SQLite, MS-SQL, GCP, AWS, Datadog, Tinybird, Grafana, Svelte, React, Runpod.

Some Proof : My website is written in Svelte (https://river.berlin) I used to work at Runpod previously, mostly on MLOps, I love training models, or doing the slow work of figuring out vllm/sglang issues. I used to teach python an incredibly long time ago (https://www.youtube.com/@typecodelearn6971) obviously my skills have increased considerably since then. I like deep slow work, things that take a few days to a few hours to solve - That being said I am open to everything, if you want it made, I probably can make it.

Résumé/CV: See it on my website : https://river.berlin

Email : [email protected] (or alternatively use the form on my website)

I've 7 years of experience - I started of building services for individuals, and then moved on to ML-related work. I worked at RunPod last, directly working with a number of clients for runpod (including OpenAI, Cursor, Perplexity).

в прошлом месяцеdiscuss

I believe despite quantisation they were still extremely fast, which is still incredibly useful if you don't need high precision/accuracy (which is good enough for many use cases)

в прошлом месяцеdiscuss

This is super interesting, I'm particularly excited for this one as it may allow teams to scale this architecture for VLAs (vision language action models), and having sparser models means more real-time actions on a locally hosted model

demo link for anyone that wants to try this out https://playground.liquid.ai/chat?model=cmppnbgse000004l4bc8...

2 месяца назадdiscuss

Curious can you elaborate on this?

2 месяца назадdiscuss

I used their online api, and asked it to create code for a timer i can copy paste into about:blank to test out (prompt below)

it did it successfully, but it did need a follow up correction prompt, overally pretty impressive for a model with 760M active parameters, but definitely not deepseek-r1 level

that being said, if something with 760M active parameters can be this good then, there's a good chance it is likely that api-based models are likely to get cheaper in the future

Prompt ------

``` can you write me some js code (that i can put in the console for about:blank) which will basically create a timer for for me that i can start, stop, and store current values for (or rather lap)

so i want it to create buttons (start, stop, lap buttons) on the page for me with labels and divs and other elements that accordingly record the current information and display the current information, and can accordingly start, stop and lap :)

the js code that i copy paste automatically creates the html buttons and divs and other elements that can manage the timer and accordingly the timer works with them ```

2 месяца назадdiscuss

I don't think it's an inferiority complex, negativity sells more and carefully understanding things doesn't sell as much

3 месяца назадdiscuss

If you look at their whitepaper (https://github.com/PrismML-Eng/Bonsai-demo/blob/main/1-bit-b...) you'll notice that it does have some tradeoffs due to model intelligence being reduced (page 10)

The average of MMLU Redux,MuSR,GSM8K,Human Eval+,IFEval,BFCLv3 for this model is 70.5 compared to 79.3 for Qwen3, that being said the model is also having a 16x smaller size and is 6x faster on a 4090....so it is a tradeoff that is pretty respectable

I'd be interested in fine tuning code here personally

3 месяца назадdiscuss

The link didn't work for me personally, but that may be a bandwidth issue with me fighting for a connection in the EU

3 месяца назадdiscuss

here's the google colab link, https://colab.research.google.com/drive/1EzyAaQ2nwDv_1X0jaC5... since the ngrok like likely got ddosed by the number of individuals coming along

5 месяцев назадdiscuss

This depends on how much better the models will get from now in, if Claude Opus 4.6 was transformed into one of these chips and ran at a hypothetical 17k tokens/second, I'm sure that would be astounding, this depends on how much better claude Opus 5 would be compared to the current generation

5 месяцев назадdiscuss

Robotics benefits nearly every industry dependent on robotics, being able to expand your market base to countries where individuals have lower incomes obviously provides a boost to companies that wish to sell more product.

It also allows for companies to reclaim their supply chains within the country of manufacture (for high income countries)

5 месяцев назадdiscuss

I just got 50 emails lol, this really sucks, phew glad i am not alone

6 месяцев назадdiscuss

The filters don't actually work, unless I'm misunderstanding them >.<, can you fix them please?

7 месяцев назадdiscuss

yes!, thanks for the link!

7 месяцев назадdiscuss

yup that's what I meant!, Jevon's paradox applies to resource usage in general and not towards a specific companies dominance

if computational efficiency goes up (thanks for the correction), and CPU inference becomes viable for most practical applications, GPUs (or accelerators) themselves may be unnecessary for most practical functions

7 месяцев назадdiscuss

Due to perverse incentives and the historical nature of models over-claiming accuracy, it's very hard to believe anything until it is open source and can be tested out

that being said, I do very much believe that computational efficiency of models is going to go up [correction] drastically over the coming months, which does pose interesting questions over nvidia's throne

*previously miswrote and said computational efficiency will go down

7 месяцев назадdiscuss

it's fine now...I believe