Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: Predictions for 2025?

news.ycombinator.com

Post reply on HN
Print view View original

331–340 of 620 posts

Re: Ask HN: Predictions for 2025?

#331

Post by fnqi8ckfek »

doku wrote: post #271

90% confidence level, Hottest weather in recorded history. Extreme weather in many places not seen before. No one is prepared for this commonplace outlier, Globally common, locally outlier. Huge insurance payout but most not fullfilled. 90% CL: Top search engine results completely inundated by LLM text, we none the wiser. SEO solved. Search engine switches to new algorithm based on LLM. 90% CL: AI with tool use that…

How do those confidence levels work? It looks to me you're just assigning probabilities to outcomes.

Re: Ask HN: Predictions for 2025?

#332

Post by rakkhi »

doku wrote: post #271

90% confidence level, Hottest weather in recorded history. Extreme weather in many places not seen before. No one is prepared for this commonplace outlier, Globally common, locally outlier. Huge insurance payout but most not fullfilled. 90% CL: Top search engine results completely inundated by LLM text, we none the wiser. SEO solved. Search engine switches to new algorithm based on LLM. 90% CL: AI with tool use that…

I did my top 5 for my field cyber security

AI in cyber defence and attack Quantum resistant cryptography Death of the password? Passwordless, passkeys, phishing resistant MFA More suppliers compromised, another major open source project long con Doing more with less: automation and consolidation

https://rakkhi.substack.com/p/cyber-security-predictions-for...

Re: Ask HN: Predictions for 2025?

#333

Post by twometwo »

As a lazy reader, could I ask those posters that use linear models for prediction to just say so, for example use tags as: [Linear prediction based on the last five years for bitcoin value]. Since I can make easily that prediction by myself I am not going to extract much value with your linear prediction. Anyway that prediction reflects what you are interested in or what is interesting for you, or what you consider could be interesting for the rest of us. That is the answer to another question: What do you think is interesting for us.

Re: Ask HN: Predictions for 2025?

#334

Post by stogot »

rakkhi wrote: post #332
doku wrote: post #271

90% confidence level, Hottest weather in recorded history. Extreme weather in many places not seen before. No one is prepared for this commonplace outlier, Globally common, locally outlier. Huge insurance payout but most not fullfilled. 90% CL: Top search engine results completely inundated by LLM text, we none the wiser. SEO solved. Search engine switches to new algorithm based on LLM. 90% CL: AI with tool use that…

I did my top 5 for my field cyber security AI in cyber defence and attack Quantum resistant cryptography Death of the password? Passwordless, passkeys, phishing resistant MFA More suppliers compromised, another major open source project long con Doing more with less: automation and consolidation https://rakkhi.substack.com/p/cyber-security-predictions-for...

Those are generic enough you could argue they already occurred last year or this year?

Re: Ask HN: Predictions for 2025?

#335

Post by wepple »

mranderson_neo wrote: post #327

AI will wipe out 80% of jobs. People will starve for employment. Unemployment rates will be highest in History, even more than war times. There will be a general purposelessness among youngsters because of this. And it will create a lot of social issues and unrest in the world

You don’t understand what most people do for work. I’d be very surprised if 10% of jobs are wiped out.

Re: Ask HN: Predictions for 2025?

#336

Post by rakkhi »

stogot wrote: post #334
rakkhi wrote: post #332

Earlier quoted context omitted.

I did my top 5 for my field cyber security AI in cyber defence and attack Quantum resistant cryptography Death of the password? Passwordless, passkeys, phishing resistant MFA More suppliers compromised, another major open source project long con Doing more with less: automation and consolidation https://rakkhi.substack.com/p/cyber-security-predictions-for...

Those are generic enough you could argue they already occurred last year or this year?

There is a lot more detail in the article that makes not generic

Re: Ask HN: Predictions for 2025?

#337

Post by wepple »

jiraiya0 wrote: post #238

- More widespread use of AI in everyday applications - Hybrid work models become the standard. More companies embrace "digital-first" policies - More AR/VR applications in daily life - Lab-grown meat and plant-based protein alternatives will become more mainstream - Investment and adoption of renewable energy technologies like solar, wind, and geothermal will accelerate - Further development of CRISPR and other gene…

Most of your points feel very “water discovered to be wet”

Except: Lab-grown meat and plant-based protein alternatives will become more mainstream

That one seems to have been trending negatively and I expect it to continue

Re: Ask HN: Predictions for 2025?

#338

Post by tonyedgecombe »

mytailorisrich wrote: post #326
dachris wrote: post #264

Industrial civilization will continue to follow along the "Business as Usual" trajectory of the 1972 Limits to Growth publication [0]. [0] https://en.wikipedia.org/wiki/The_Limits_to_Growth#Compariso...

" is a 1972 report that discussed the possibility of exponential economic and population growth with finite supply of resources " The main problem is that for many reasons people tend to focus on the economic growth part while the population growth part is the killer.

Population growth is a solved problem. We have already passed peak births.

Re: Ask HN: Predictions for 2025?

#339

Post by hypoxia »

OutOfHere wrote: post #106
hypoxia wrote: post #33

I think the defining story of 2025 will be AI agents getting very good with computer use, largely enabled by RL fine tuning.

Could you help understand the importance of RL finetuning? What can it accomplish that regular finetuning can't? What's a use case for it?

From my experience there are three key issues with agents today:

1. They usually don't end up completing the right set of steps required to complete tasks when using our human-defined frameworks (react, rewoo, supervisor-worker, teams of multi-agents, etc.)

2. They get lost easily, and forget what they were doing or complete the same tasks over and over in a loop (bad planning)

3. They exit early, thinking they have completed the task when they have not (bad evaluation)

The jump in reasoning ability from 4o to o3 will enable a drastic improvement in planning and execution within our human defined frameworks.

But, more importantly, I believe RL fine tuning will enable the model to learn better general approaches to planning and executing steps to complete work. This is Sutton's bitter lesson at work.

For me, desktop automation is the killer app of RL fine tuning, rather than better reasoning in chatbot apps and APIs.

When OpenAI releases their desktop agent capabilities built on this, hopefully in Jan, I think we're going to see another ChatGPT moment.

Even if not, the ability to easily train the system to complete your tasks successfully with full desktop usage is going to be a major unlock for enterprises.

More on RL fine tuning here: https://openai.com/form/rft-research-program/

Post reply on HN