I've trained a GPT on all of my company's documented use cases which is helping me with sales-acumen and copywriting. Definitely seeing it going the direction of smaller language models for real world use cases.
It’s because these news outlets just need to drum up clicks. They ignore the stuff that actually matters to pay their bills. Is what it is not hate i just tune it out.
How is this peaking? Have you tried the latest models released in 2024? They're very good. Even if we see a 2 fold increase in reasoning in 2025, it's still going to be smarter and more capable than most people on Earth.
Please tell us the impact on education this year. (I am a teacher of Physics and would love to hear your thgouhts on the impact on our students.) Thank you - lovin your reporting xx
She is wrong. Three components of Reasoning for AI: 1. Foundation (Pre-training) 2. Self-improvement (RL) 3. Test-time compute (planning). So scaling of frontier models is not going away anytime soon.
The big things to look for in 2025 are: 1. Androids - recent advancements in training have made this a very achievable goal in 2025 for less than the cost of a car. 2. Infinite context - this is going to be critical for allowing AI to do things like displace human workers since it will need to be able to remember what its job parameters are. 3. Agents - we already have them, but they are too infantile to be useful. In 2025, AI companies are focusing on this feature as it will allow AI to start doing the job of a human. Ultimately, this is where the real AI investment pays off, when AI can either replace or significantly augment human workers. We definitely have not peaked with AI, they simply hit a wall when it came to scaling. Already though China has found a way around the scaling problem, and I'm sure we'll deal with it too. They were just going after the low hanging fruit by scaling constantly. Now they will have to work smarter, not harder.
Great report. The only thing i disagree about Small LMs. They proved to be worse than LLM even in highly specilized tasks. The trends is that LLMs become faster and can run locally making SLM usless. While RAG and ever increasing content window are solving context problem. Even LLm fine tuning is questuionable. Imho.
lol are we still questioning whether ai progress has stalled when progress keeps accelerating. o3 just saturated arc-agi a benchmark that proved ai-proof for 5+ years. the next benchmark to beat is frontier math which o3 scored 25%. the arc-prize fondation announced that they are making a new benchmark in which humans will score >95% and o3
My Ai-Agent will have a PhD in inverse design. In material science these structures are based on desired properties. So apparently, some humans haven't been using their PhD at an optimal standard. How much desire will that make for me.
When are people going to get a clue and realize you can’t just make an AI that is more advanced. The way it really works is that the current AI continues to advance over time. That’s why ChatGPT will continue to be on top because it was the first.
Inferencing also depends on, on hand product good luck getting enough TPUs to build a data center. Security is more likely to be the concern in 2025 as foreign agents look to duplicate advanced AI systems for less.
Thanks for the continued updates. I have given up on stocks, my money is melting day by day. Crypto currency and Bitcoin ETFs will outsmart the banking system in the nearest future serving as a global fiat. Already making over 90% profit from my current investment.
Right now Bitcoin has alot of opportunities, much more than other financial instruments. I am glad I got into crypto when I did. With Bitcoin ETFs, from $17K to $45K that's the minimum range of profit return weekly regardless of how bad it gets on the economy
Ms Charlotte Patton, holds the credit of being crucial about crypto and stocks trading and also a knowledgeable Brooker. Serves as my family's personal broker, as well as to many families across the United States. As well licensed as a FINRA agent in the U.S.
@amols101 I know.. it was a general comment, as you see Nvidia price dip of 10-15 % now a days. And it is cheaper thsn AVGO, Amazon, Apple, Tesla, etc.
It was stupid for these big companies to open source their AI software. Everyone can create these AI models now. It also invites and makes it easy for governments to regulate them and control how they work.
I've trained a GPT on all of my company's documented use cases which is helping me with sales-acumen and copywriting. Definitely seeing it going the direction of smaller language models for real world use cases.
The fact that this question can be raised after the o3 benchmarks is crazy to me.
I second that.
More Chinese knock offs. They can’t create so they steal, and make it cheaper, with less quality.
I third that.
It’s because these news outlets just need to drum up clicks. They ignore the stuff that actually matters to pay their bills. Is what it is not hate i just tune it out.
Dierdre Bosa! 🎉🎉🎉great report! Thanks!
How is this peaking? Have you tried the latest models released in 2024? They're very good. Even if we see a 2 fold increase in reasoning in 2025, it's still going to be smarter and more capable than most people on Earth.
1:29 “western model”😂 that killed me😂😂😂
Why? AI is being developed in the west too
Please tell us the impact on education this year. (I am a teacher of Physics and would love to hear your thgouhts on the impact on our students.)
Thank you - lovin your reporting xx
thanks dierdre bosa! happy nye and best wishes to u and urs in 2025 📈🇺🇸
💚🖤
She is wrong. Three components of Reasoning for AI:
1. Foundation (Pre-training)
2. Self-improvement (RL)
3. Test-time compute (planning).
So scaling of frontier models is not going away anytime soon.
The big things to look for in 2025 are:
1. Androids - recent advancements in training have made this a very achievable goal in 2025 for less than the cost of a car.
2. Infinite context - this is going to be critical for allowing AI to do things like displace human workers since it will need to be able to remember what its job parameters are.
3. Agents - we already have them, but they are too infantile to be useful. In 2025, AI companies are focusing on this feature as it will allow AI to start doing the job of a human. Ultimately, this is where the real AI investment pays off, when AI can either replace or significantly augment human workers.
We definitely have not peaked with AI, they simply hit a wall when it came to scaling. Already though China has found a way around the scaling problem, and I'm sure we'll deal with it too. They were just going after the low hanging fruit by scaling constantly. Now they will have to work smarter, not harder.
i created chat bot for food ordering how i can convert to AI agent
Great report. The only thing i disagree about Small LMs. They proved to be worse than LLM even in highly specilized tasks. The trends is that LLMs become faster and can run locally making SLM usless. While RAG and ever increasing content window are solving context problem. Even LLm fine tuning is questuionable. Imho.
Sorry D, it’s not about smaller models, it’s about agents
a lot of use cases are impossible because of latency but yes agents will be the big step forward in use cases
when one android is more effective than 100 of the best trained agents, soldiers and policemen combined
then you all understand
lol are we still questioning whether ai progress has stalled when progress keeps accelerating. o3 just saturated arc-agi a benchmark that proved ai-proof for 5+ years. the next benchmark to beat is frontier math which o3 scored 25%. the arc-prize fondation announced that they are making a new benchmark in which humans will score >95% and o3
Elsewhere is in the upswing? Cuz elsewhere they are just starting.
good stuff, more please
FYI, all cloud providers have agents, not only Microsoft. I'm using AWS Bedrock.
Deepseek is real open ai. OpenAL is closed AI
Basically no insight here, just dropping names to sound like they know what’s going on
My Ai-Agent will have a PhD in inverse design. In material science these structures are based on desired properties.
So apparently, some humans haven't been using their PhD at an optimal standard. How much desire will that make for me.
When are people going to get a clue and realize you can’t just make an AI that is more advanced. The way it really works is that the current AI continues to advance over time. That’s why ChatGPT will continue to be on top because it was the first.
TPU and ASICs have limited inferencing capabilities. Can not be programmed. Most of inferencing will have to be done in the Cloud. Just searches.
Inferencing also depends on, on hand product good luck getting enough TPUs to build a data center. Security is more likely to be the concern in 2025 as foreign agents look to duplicate advanced AI systems for less.
take on Philips, IBM, Microsoft, they should know the truth about me and the patent without which this technology would not exist
Well, Deirdre Bosa, wrong yet again.
In other news, the Sun is up today.
Launching our news platform this week. Why don’t you all interview me lol.
Who will be the Prodigy or AOL of AI, because I want to invest in them😅
is Dierdre Bosa the chuckle brother's sister ?
Thanks for the continued updates. I have given up on stocks, my money is melting day by day. Crypto currency and Bitcoin ETFs will outsmart the banking system in the nearest future serving as a
global fiat. Already making over 90% profit from my current investment.
Right now Bitcoin has alot of opportunities, much more than other financial instruments. I am glad I got into crypto when I did. With Bitcoin ETFs, from $17K to $45K that's the minimum range of profit return weekly regardless of how bad it gets on the economy
I've invested in Bitcoin, Forex and stocks. I'm not really happy with what's going on. Can you advise me on what to do?
Get a professional to guide you.
Ms Charlotte Patton, holds the credit of being crucial about crypto and stocks trading and also a knowledgeable Brooker. Serves as my family's personal broker, as well as to many families across the United States. As well licensed as a FINRA agent in the U.S.
I'd recommend Charlotte Patton too, we started my investment with $7,000, a week later, we had grown to $25,000. This woman is an amazing genius
Tesla is the real use case
At this moment I think reasoning is not real
Ai didn't come close to peaking in 2024. Palantir is the company to be in. They've been having used cases in all 2024, the rest are all behind.
Tesla is the only ai company .
again and again, bunch of biased BS !!
AI hasn't peaked. Not at all. What an uninformed broadcaster.
Indian are good at PPT. Let's get real engineers
You wanna keep betting against NVIDIA? OK, KEEP doing that! Doesn't matter to NVIDIA
She didn't say that.
@amols101 I know.. it was a general comment, as you see Nvidia price dip of 10-15 % now a days. And it is cheaper thsn AVGO, Amazon, Apple, Tesla, etc.
The US needs to issue more H1B visas to Indians, they are good at JavaScript,...etc.,
The race for AI supremacy is heating up! With new players entering the field, it's anyone's game. Who are you betting on to be the 'Google' of AI?
I love Dierdre Bosa, i want to ask her out on a date. ❤
ok, but don't ask for investment advice!
and i'm sure you're exactly what she's looking for too
you should definitely try. Im sure shes waiting for you to ask her.
@anzatzi,
FACTS! She's Wrong 99.9% of the time, especially with AI.
@@otherZinc she's a reporter, she doesn't do stock picks. you new here?
It was stupid for these big companies to open source their AI software. Everyone can create these AI models now. It also invites and makes it easy for governments to regulate them and control how they work.
too outdated, can be open resource
Open source models have some limitations though