ramoz

31 ms·

631.
by ramoz 2y ago
He said nothing controversial and yes the model has digressed. This "flagship model" is fast and cheaper, but it is worse (regardless of benchmarks & charts, subjectivity is failing for me and many others). The model seems bet
632.
by ramoz 2y ago
From: Is building intelligent machines a good idea? > If one accepts that the impact of truly intelligent machines is likely to be profound, and that there is at least a small probability of this happening in the foreseeable future, it i
633.
by ramoz 2y ago
Well here’s a somewhat limited version of your idea and really only helps mitigate the copy/paste effort with coding: https://github.com/backnotprop/prompt-tower My original idea was a DnD interface that works at
634.
by ramoz 2y ago
I like this and wished openai or anthropic enabled similar in their UIs... it would be simple actually: "create a new chat from here" Otherwise, great job! It's cool, but it's pricey and that is a personal deterrence.
635.
Prompt Tower v0.1.9 – Batch file selection(github.com)
3 points by ramoz 2y ago|0 comments
636.
Show HN: Prompt Tower – Simplify writing prompts with multiple code blocks
10 points by ramoz 2y ago|0 comments
637.
by ramoz 2y ago
Similar, but life hit me fairly hard in my early thirties. Things can and will go wrong for all of us. Being in a pit of actual negativity is an interesting place to be. Which led me to mindfulness practice and understanding what "the
638.
by ramoz 3y ago
My anecdote: I assume I have fairly severe tinnitus at this point. It was in part exposure to loud noises over the years (military), but I think chemotherapy really did me in the past year as tinnitus has become more bothersome than ever; c
639.
by ramoz 3y ago
A lot of people work in tech these days. A lot of people write blogs. Anyway... back to scale, consistency, and speed - IAC.
640.
by ramoz 3y ago
I’ve worked in scaled enterprise search, both with lexical (lucene based, eg elastic search) & semantic search engines (vector retrieval). Vector retrieval that isn’t contextualized in the domain is usually bad (RAG solutions call this
641.
by ramoz 3y ago
Depending on the application, you would do continued pretraining over new tokens to gain new knowledge. 100M tokens is applicable here. You would fine-tune, certainly, for domain-specific tasks, and would curate a subset of the 100M tokens.
642.
by ramoz 3y ago
Their goal is to always drive enterprise business towards consumption. With AI they need to desperately steer the narrative away from API based services (OpenAI). By training LLMs, they build sales artifacts (stories, references, even accel
643.
by ramoz 3y ago
Bearish. These types of tools/agents-chaining will be irrelevant due to lackluster capability until AGI is achieved. At which point, the basis for creating these types of tools/agents will be defunct.
644.
by ramoz 3y ago
As someone who has built large scale AI, it’s funny to see the only 2 brand name tools my team used missing: Airflow (etl and embedding processing), BigQuery (product analytics)
645.
by ramoz 3y ago
https://codepen.io/backnotprop/pen/qBvzpvr
646.
by ramoz 3y ago
<div style="display:flex;justify-content:center;align-items:center;"> <div>I'm centered</div> </div> * I wrote the original without testing but it is justify-content, not justify-item
647.
by ramoz 3y ago
Yea… agree. I’m not resonating with the downvotes here on similar comments. ChatGPT goes above and beyond for me in many ways. Tests seem… easy in terms of gpt capabilities. Last week I had it write python that traversed an AST and construc
648.
by ramoz 3y ago
Your methodology is nice. Is there any work or planned work around enterprise authentication and access? For instance, indexing Sharepoint in such a way where a user of Danswer isn't exposed to sharepoint information they wouldn'
649.
by ramoz 3y ago
Here is a fourth: https://x.com/james_e_seale/status/1760348535608725716?s=46&...
650.
by ramoz 3y ago
I was wondering if these models would perform in such a way, given this week's X/twitter storm over Gemini generated images. E.g. https://x.com/debarghya_das/status/1759786243519615169?s=20 https:/
651.
by ramoz 3y ago
> Scaling Inference We additionally scale our inference code to support million-length sequences by implementing RingAttention for decoding. Inference for such long sequences requires a minimum of v4-128 with a TPU mesh sharding of 32 te
652.
by ramoz 3y ago
for the uninitiated: <div style="display:flex;justify-items:center;align-items:center"> <div>I'm centered</div> </div>
653.
by ramoz 3y ago
https://chat.openai.com/share/c89aaac5-a6f7-4f50-9cac-ae4777...
654.
by ramoz 3y ago
For me, GPT nailed it with this prompt: Tabitha likes cookies but not cake. She likes mutton but not lamb, and she likes okra but not squash. Following this rule, will she like cherries or pears? Take a deep breath in order to solve the pro
655.
by ramoz 3y ago
Do you work in the medical industry? (To be Frank it sounds like you’re spewing bs about a topic you have no intimate knowledge of - seems rather naive actually).
656.
by ramoz 3y ago
I mean it depends on your workloads. Any engineer faced with deploying AI systems at scale will inevitably break through abstractions (or fail). AI is further proof that our systems aren’t simply getting easier to build around.
657.
by ramoz 3y ago
There’s no ocr or ai involved here (other than the standard fallback). What this library, and something like fitz/pymupdf, allow you to do is extract the text straight from the pdf, using rules about how to parse & structure it. (M
658.
by ramoz 3y ago
For me, PyMuPDF/fitz has been the best way to retain natural reading order and set dynamic enough rules to extract text in complex layouts. None of the mentioned tools did this out of the box, none seemed easy to configured, all defini
659.
by ramoz 3y ago
My team is trying to get to that, albeit are very early and hardly integrated enough into training processes/scaled-inference https://huggingface.co/spaces/EQTYLab/lineage-explorer
660.
Show HN: A lineage explorer for open source models and datasets(huggingface.co)
3 points by ramoz 3y ago|0 comments

More ›