Weekend Food for thought - Artificial Intelligence

[Konebhar6]

9 hours ago, paaparao said:

appatlo big data, data science ani paiki leparu. avi paththa lekunda poyayi. 

ipudu AI.  

 

Bid data and data science are the core of any company now. They are using analytics to drive the company forward.

Too many graduates coming into market with this skill set though. 1000 graduates from Berkeley itself.

look at all the big data companies flourishing. Snowflake, mangoDB, data bricks, etc.

[paaparao]

53 minutes ago, Konebhar6 said:

Bid data and data science are the core of any company now. They are using analytics to drive the company forward.

Too many graduates coming into market with this skill set though. 1000 graduates from Berkeley itself.

look at all the big data companies flourishing. Snowflake, mangoDB, data bricks, etc.

good to know. they became SaaS type service to companies and it is healthy. not like they promoted as it will replace many workers.

[kevinUsa]

1 hour ago, Konebhar6 said:

Bid data and data science are the core of any company now. They are using analytics to drive the company forward.

Too many graduates coming into market with this skill set though. 1000 graduates from Berkeley itself.

look at all the big data companies flourishing. Snowflake, mangoDB, data bricks, etc.

What Abt sap anna

[Konebhar6]

7 minutes ago, kevinUsa said:

What Abt sap anna

ERP systems are back bone for business operations. They will stay. All ERP's will do good for future. 

[kevinUsa]

20 minutes ago, Konebhar6 said:

ERP systems are back bone for business operations. They will stay. All ERP's will do good for future. 

Acumatica ani amazon di undi better ee antava

[Konebhar6]

14 minutes ago, kevinUsa said:

Acumatica ani amazon di undi better ee antava

No idea on it. SAP, Oracle are most popular. After that Workday is doing really since last few years.

[ramannar]

31 minutes ago, Konebhar6 said:

No idea on it. SAP, Oracle are most popular. After that Workday is doing really since last few years.

Workday similar to oracle hcm? Even that they want to inject gen ai

[Konebhar6]

25 minutes ago, ramannar said:

Workday similar to oracle hcm? Even that they want to inject gen ai

Workday has a complete ERP system. It was more popular for HCM before, but they have complete suite and a lot of clients are using. But SAP, Oracle still leaders. 

Gen AI is being loosely thrown around. Really have to see how its being used in the context. 

iyyalrepu, even API's are being marketed as AI. 

[ramannar]

1 minute ago, Konebhar6 said:

Workday has a complete ERP system. It was more popular for HCM before, but they have complete suite and a lot of clients are using. But SAP, Oracle still leaders. 

Gen AI is being loosely thrown around. Really have to see how its being used in the context. 

iyyalrepu, even API's are being marketed as AI. 

esg wise big hole annaw still leaders mestri comfanies and witch fellas overacting ai narrative everywhere

[Konebhar6]

1 minute ago, ramannar said:

esg wise big hole annaw still leaders mestri comfanies and witch fellas overacting ai narrative everywhere

Some companies will come up with good AI solutions in future. But data labelling is super expensive. It needs to be done by professionals in that particular space. For e.g. if an app is built to give Legal solution, the data needs to be tagged properly and the authenticity of results need to be validated by professionals. Its not cheap. 

A lot of companies will get funding, they will also close shops unable to close shops (leadership is critical here). But some good solutions will come out.

In the mean while, everyone is using the AI term without really understanding what it is. Mestri companies gurinchi inka cheppakarledu.

[kevinUsa]

3 hours ago, Konebhar6 said:

Some companies will come up with good AI solutions in future. But data labelling is super expensive. It needs to be done by professionals in that particular space. For e.g. if an app is built to give Legal solution, the data needs to be tagged properly and the authenticity of results need to be validated by professionals. Its not cheap. 

A lot of companies will get funding, they will also close shops unable to close shops (leadership is critical here). But some good solutions will come out.

In the mean while, everyone is using the AI term without really understanding what it is. Mestri companies gurinchi inka cheppakarledu.

Ee data labelling kin course ekkada untadi ??

[Thokkalee]

14 hours ago, Konebhar6 said:

I never understood Blockchain. I spent a lot of time. Talked to people. No one could explain properly. So imagine the plight of non-IT people.

AI is different. Who thought Self-driving cars were possible. But we have them. Its a different story whether people are buying them or not. May be some practical issues with them.

AI is still very primitive. Even after 5 yrs it will be primitive. But it will change the way we do certain things. 

I dont think it will replace everything. It will fail in a lot of places, but a lot of good use-cases will come out of it.

What made AI/ML applications more usable and more prominent is the emergence of cloud tech and also the new processors that can train the models.. my engineering project is based on pattern recognition, which was long time ago.. 

okko area thiskuni models train chestaaru.. it may take years for each model to become perfect (2 9’s to 11 9’s).. but there will still be deficiencies in the models.. Tesla has been training their autonomous driving software with data of over a few billion miles of driving data.. still it is not perfect.. 

[Konebhar6]

2 hours ago, kevinUsa said:

Ee data labelling kin course ekkada untadi ??

Oka sari nenu pettina yahoo link chudu. Not everyone can work on data labeling. 

[Thokkalee]

 

[pichukgudu]

On 8/24/2024 at 6:19 AM, Konebhar6 said:

We all did a lot of Disco here discussing AI/ML and how they are going to take our jobs. See the below Link

https://finance.yahoo.com/news/hidden-reason-ai-costs-soaring-174705926.html

Interesting points from the article

• Rising costs for AI companies due to Data Labelling. This is the process of creating labels/tags for data so that the model can recognize and interpret patterns
• Data Labelling has so far been used to develop AI models for self-driving cars to recognize images of pedestrians, trucks, and stop signs labelled properly. 
• LLMs go through Reinforcement Learning Human Feedback where humans provide qualitative feedback or ranking on what the model produces.
• Data Labelling companies like Scale AI, Snorkel AI, Otherside AI, CareYaya, etc received a lot of funding but confirmed rising costs. DON'T RUSH TO LOOK FOR THE STOCK CODES. THESE ARE NOT LISTED. THERE YOU GO, I SAVED YOU SOME TIME. 

Why Costs are going up - They have to now hire professionals or highly technical people in fields like Finance, Legal and Health Care. Basically, Doctors, Lawyers, PHDs, etc

Its just funny, how when mouths to feed are growing, by each passing day .. capitalists talk about replacing humans by machines as part of reducing investment at some point, which in my opinion may not sustain beyond a level!

Also, no matter how well you train a machine to replicate a human behaviour, its always vulnerable to confusion in ambiguous detailing of things!

As its just in primitive stage .. there is a lot of raving about its capabilities .. and we'll find out soon the reality of it.