o1-Preview: 11 STUNNING Use Cases

I have a comment I am sure everybody watching your otherwise excellent video can agree on:
1. The video-length you might have saved by a/ Cutting the sentences before they were finished. b/ using speeded up variable speed staccato speech, is totally offset by the collateral reduction of viewer experience quality. It is extremely uncomfortable to be
subjected to the pre-sentence-finishing cuts, and this extra sentence- finishing-guessing -work you consequently are subjecting us viewers to, is both unnececessary and counterproductive. Slow down! Re-edit! 😊

The AI hype train has arrived again…. we're being bombarded with words like "insane", "incredible" and "amazing" again.

>misdiagnoses aren't the doctor's fault
They literally are. What is this zero responsibility culture we find ourselves in?

You would like a unique business idea?

Here it is.

A modular adapter for interchangeable tool use for humanoid robots.

To use generative technologies to assess the strengths of a modular component interface that can provide a stable interchangeable multifunctional platform for tool use to provide multiple engineering capabilities for a standardisation and modularisation to be used to give robots multiple tool options from types of rotary interfacing, hydraulic interfacing, pneumatic or even impact related.

The Ai can be used to assess both tensile, shear, yield and compressive strengths of the component parts using structural and material science data. Viable volumetric data and flow calculations for running different types of rotary tooling from cutting, abrasive, reaming, milling and turning. Pneumatic for use also with rotary tool platforms as well as for cleaning and painting, maybe to supply air to stranded humans trapped in confined spaces where other humans may not be able to easily of safely gain access because of incapacitating conditions and other existing pneumatic platforms that already exist. Hydraulic for shearing, folding and bending as well as for emergency service use by fire services needing to extract people trapped in potentially multiple situations. The impact tool can be used both with air actuated chiselling, scabling, as well as in combination with rotary for torque related fastening of fixings.

Using both reasoning platforms and image generation with calibration for engineering quality drawings providing a design that can be adopted by all industry based robotics platforms.

They already know we yearning for (QET) & (NMR).They have proved Hotta's theory. All you fossil founders on this planet!They can get that energy & it doesn't violate physical law. My hype is entangled in quantumness. Imagination is more powerful than knowledge.

Sharing your health and legal issues with a soulless company that is not held to the same standards as real doctors and lawyers? What a great idea! /s

Why people only care about the coding part of LLMs? What about the mathematics and physics it can do?

They are just doing what open source has been doing. Asking the model to think through steps is not new and not a model.

OpenAI's o1 is a significant advancement in generative AI. It has numerous applications for individuals, although it is not intended to replace jobs like developers, lawyers, and doctors. Furthermore, it lacks internet access, and if that were to occur, it would be a game changer.

I dont hear any audio from this video. :/

I've subscribed but the more hype the less likes

Hmm.. Did you ask the model "What do you think about me"? Most likely o1 will hide behind the censorship layer and will answer "As an LLM I am unable to…". You need a DAN script to discover what the model really thinks about you 🙂. Good luck.

This such a scam it is hard to watch how people are dooled by this.

Syllables being lost? I've greatly appreciated your videos about the use of frontier AI systems to solve real-world problems. Unfortunately, they sometimes are hard to understand due to the loss of parts of the ending syllables from words at the ends of phrases and sentences, in addition to the loss of the very brief pauses which normally would be present between sentences. This current video seems to suffer in all of those regards, especially in its last section (the one about legal advice) which occasionally seems to lose entire words. Is there any way this situation could be remedied? Thanks for whatever help you can provide.

Be carefull, GPT is sure of it's answers and yesterday it mixed things. After pointing reference to it, it changes it's answers. So be cautious.

I used it in order to make some changes to an already complex Android app and, honestly, if you don't have code skills, programming logic, prompt knowledge, software engineering knowledge (including requisites redaction) and project understanding, it will be useless. If you don't know to prune the answer the machine gives, it'll generate a vicious circle of coding errors. It is amazing, but there's also a lot of hype and exaggeration towards it.

I found the coding output from o1 to be slightly worse than Claude Sonnet 3.5. has anyone else done a comparison?

It just stinks how restricted the uses will be. I thought about paying the $20 a month when 4o came out but I couldnt even figure out how many uses that would get me and I just decided to pass. 3 & 4 with the occasional 4o do just fine.

You would think with what I just said that people shouldnt complain with whatever the number of GPT 1o uses they get are, but when I can only use the model 4 times a day I will not really think to incorporate it into my workflow. Maybe on special occasions, im sure it will be a nice once in a while gimmick to try out, but I couldn't really see it revolutionizing my business or impacting my life in any way.

anyone can tell me how can I find the exact background of the first few minutes of video for editing, I truly appreciate anyone can help me please

we need a stylish o1 ear to ear brainwave detector with stereo ear bud cameras and audio for real time 360 degree world context every time you ask sub vocally how many "R"s in strawberry..

Just tested with the o1 beta on a space module concept. I have tried it with claude 3.5 and grok and chatgpt 3 and 4 .. they all kind of get it but only after many prompts and corrections and recalculation. never really getting it correct and grok gets really confused. BUT ! o1 !!! got it with only a couple of clarifications. (my mistakes for not being clear and specific) !! I am very impressed.. sure it takes a while to analyse and process.. but It got the answer !! in about 2-5% of the time and with much less effort on my part !! PS I tried shapiro's prompt simulation of o1 applied on Grok 2 also helped a lot !! he is on to something !!

I almost did a spit take when he said, "the editor of 2024. Vim"

RIGHT ON BROTHER!! Vim is the editor of ALL TIMES!!! (Notepad++ is pretty good to, it's invited to the after party 🙃)