nmitchko

Poster: The cool part is that the model literally talks itself out of giving the cached wrong answer with a latent only thinking budget.

A fantastically simple solution to improving algorithms, I wish I had this years ago in activation engineering: https://blog.n.ichol.ai/llm-activation-engineering-an-easy-f...

How do I access AlphaEvolve?

I used those two in combination to fix pain after 3x surgeries to repair a torn pec + infection. They work and helped me heal from being at a 3/10 constant pain down to baseline.

Not something I would do at any point for fun. But anecdotally, it's materially better than other alternatives offered/available.

Can someone make a startup that allows me to do this as an individual?

In case anyone wants to do this themselves, check out the pipeline here: https://github.com/isc-nmitchko/iris-document-search

Colnomic and nvidia models are great for embedding images and MUVERA can transform those to 1D vectors.

Next steps for AI in general:

  - additional modalities
  - Faster FPS (inferences per second)
  - Reaction time tuning (latency vs quality tradeoff) for visual and audio inputs/outputs
  - built-in planning modules in the architecture (think premotor frontal lobe)
  - time awareness during inference (towards an always inferring / always learning architecture)

No other models that are public worth comparing to... Hippocratic advertises good benchmarks but that might be marketing fluff.

Have you checked out dataset building with nemotron? The nemotron synthetic data builder is quite powerful.

Moreso, check out model merging. It's possible if you merge some of your model against llama3.1 base it may perform much better.

Check out max labonne's work on hugging face

Interesting they don't compare to open-bio. Page 7 charts are quite weak.

https://huggingface.co/aaditya/Llama3-OpenBioLLM-70B

We're excited to share pitchpilot with the HN community. Our beta users have found the embedded audio particularly useful for enterprise sharing. We're keen to keep improving, and our mission is to make communication easier.

In the roadmap is adding video export, digital twin presentations, and real-time presentations. We don't wrap a public LLM, so we don't share any data.

Given that Generative AI can now read brain scans [1] and this, I wonder how far away we are from "you thought negatively about something, the authorities are on their way".

[1] -- https://www.biorxiv.org/content/10.1101/2022.11.18.517004v3

Tin-foil hat time:

1. First, models will predict pollution. The outcomes will help shape urban policy. But these won't solve crime or stop people from driving.

2. Second, models will predict individual behavior and track person level emissions. The outcomes will force behavior changes, mostly freedom limiting.

3. Third, and finally, models will predict thoughts. The the thought of driving instead of walking might trigger a response.

It's a slippery slope and we need to draw a line between prediction and policy.

How does this compare to ehealthexchange or other qhins that have many years of experience and charge lower costs?

It truly feels like the space race in terms of building LLMs right now. Question is, who lands on the moon first?

This reminds me of the matrix movie scene when they look at the encrypted thoughts of the matrix.

https://cdn.swisscows.com/image?url=https%3A%2F%2Fi.pinimg.c...

Great work, will try this tonight.

Only question, why do you name variables with the λ symbol?

BTC =/= crypto at large.

Despite all the bad press, shady exchanges, bubbles and busts, a large amount of BTC owners don't sell, and that alone will drive it's price up. Other cryptos, I can't speak for.

California politicians confused why large companies continue to mass-migrate to Texas

Unfortunately, the NSA & NIST most likely is recommending a quantum-proof security that they've developed cryptanalysis against, either through high q-bit proprietary technology or specialized de-latticing algorithms .

The NSA is very good at math, so I'm be thoroughly surprised if this analysis was error by mistake rather than error through intent.

With all we see around cancer, it's time to study fembendazole and other anti-tubulins seriously.

For each step in the article:

1. You should really download the huggingface hosted model.

2. Why convert the model if meta already hosts it here: https://huggingface.co/meta-llama

3. This step completely glossed over the hardware requirements. Doesn't explain any of the instructions needed in the finetuning process.

4. "Now run the model". What? What about the model precision, hosting, standard open source tools, etc,etc...

This is a pretty useless post. You could also follow the same 1000x tutorials about llama and use the already uploaded hugging face formats that are on hugging face...

Here are some actually useful links

https://blog.ovhcloud.com/fine-tuning-llama-2-models-using-a...

https://huggingface.co/meta-llama/Llama-2-70b-hf

https://huggingface.co/meta-llama/Llama-2-7b-hf

Interesting that we ignore the physical environment as a contributing factor in IQ, growth, and development. For example:

* Iodine supplementation during early life (a few weeks) increases IQ * Organophosphate exposure increases autism (also why people who live near airports have higher incidence of cancer) * Phalates and non-sticks reduce male reproductive organ size * Music lessons in children increases IQ. * L-arginine during childhood increases height and free growth hormone

https://link.springer.com/article/10.1007/s11154-022-09760-7 https://www.sciencedirect.com/science/article/abs/pii/S01602... https://www.researchgate.net/publication/237236129_The_Plane...

If it ain't broke don't fix it.

You can split the model across devices with huggingface accelerate library.

Check out the infer_auto_memory_map metho which will optimize the model for your configuration (multi gpu, ram, nvme) and then run dispatch model on with that memory map.

Your understanding is correct, but I can't vouch for the claim's accuracy. This could make the execution of models much more accessible to people who don't have a 4x / RTX3090 or better in a ML or mining rig ..

As I read through this, I remind myself, there are some people that are so smart and in-depth within their field, that I have zero idea what any of the concepts they discuss are.

When I saw the title I thought the latest in security was 12 factor authentication!

This sheds a light on some bad practices around all ski areas:

* J1 / E5 visa abuse to skirt labor and wage laws and get cheap investments

* Overselling / abusive terms of carriage for lifts

* Overpricing basic food and services, assigning abusive police-style regulators on slopes (looking at you vail)

So at the end of the day, we have to take a look and think: should lifting skiers up a mountain on a chairlift really be worth billions of dollars in market cap? Should mountains, who are almost 90% public land (in the USA), even be businesses?

Time to investigate WASM-security holes for a next generation security firm

You might be underestimating the number of people who would return to the workforce for jobs like this.