Your 2024 Predictions thread

[wjfox]

Even though odds are in Trump's favor I am predicting a democrat victory for the 2024 election. I said other things in status updates a day ago but hadn't really thought about it or seen public reactions.

Kamala is normal and a good age Trump is not (like him or not you have to admit he is eccentric).
Looking at the early reactions on MAGA twitter and such there is little to no dirt on Kamala. People are treating it like a big victory to make fun of her laugh I have even seen people make fun of the fact she uses the same phrase in many speeches (like every politician ever has done).
https://x.com/libsoftiktok/status/1815112445608010006
I have seen misogynist attacks with conservative figures saying things like she has slept with people to reach the top which will only isolate them.
Much of the republican lead over dem isn't based on party policy but that they were up against someone who if wasn't famous may be in a nursing home.

My rough timeline for a democrat win would be like this.
Not much happens until the democratic convention in late august with people presuming the dems will lose.
After the convention the party will rally behind Kamala and an election platform will have been made.
Kamala and dems will slowly gain traction by appearing in the media as not only the normal candidate but as a younger and "hip" in touch one.
Plus a boost for being the potential first female president and second black president.
Celebrities such as Taylor Swift and Beyonce will start to endorse Kamala.
Trump will be told to avoid debating her but she will appear in the media and imply Trump is weak and too scared to debate her etc
He will fall for the bait due to his ego and in debates be demolished by a young normal woman who worked as a prosecutor (effectively a professional attack dog and debater).
Kamala is normal and there is little dirt on her with the opposite being true for Trump.
Unhinged attacks and conspiracies about her will only help her. She is not some super establishment figure like Clinton and this isn't the 2016 Trump who was younger and more energetic with the anti-woke pendulum swung amongst young people. (old people are usually conservative so presumably young people made the difference).

A good analysis. I think you are probably right.

[Yuli Ban]

In regards to LLMs being able to create video games, as I mentioned a GPT-4.5-class model might be able to create SNES-quality games, here's what Claude 3.5 Sonnet (which I see as a "very high end GPT-4-class" or "GPT-4.25 class" model) can already do:






There is not much distance between what's been shown here (in raw terms without overfitting like in the Minecraft example, I'd say this is easily early NES level stuff at best) and what could be feasible with a full generational leap in capabilities and reasoning.

[Powers]

In regards to LLMs being able to create video games, as I mentioned a GPT-4.5-class model might be able to create SNES-quality games, here's what Claude 3.5 Sonnet (which I see as a "very high end GPT-4-class" or "GPT-4.25 class" model) can already do:






There is not much distance between what's been shown here (in raw terms without overfitting like in the Minecraft example, I'd say this is easily early NES level stuff at best) and what could be feasible with a full generational leap in capabilities and reasoning.

Holy **** and how I haven't read about this?

[Yuli Ban]

There's a lot that LLMs can already do, that aren't really being advertised or are getting lost in the noise of the internet. Not only that, but most people use ChatGPT, and most people from that only use GPT-3.5 (or now GPT-4o) rather than any paid tier.
And even then, a lot of the capabilities of foundation models are being lost to restrictions and RLHF. GPT-4o is a fully multimodal model, but virtually none of the modalities besides text and vision have been activated. Similar with Claude if I recall. Basically, we're only using a fraction of the full power of contemporary AIs.
And that's WITHOUT including agents and tools (as they don't use these even now, outside maybe Plug-Ins and Artifacts). AI could literally be a few orders of magnitude more capable right now (as in the models already released, without further scale), but the deployment is still very early, very inefficient, and very hobbled.

[funkervogt]

Here are my 2024 predictions:

In Ukraine, the war will probably remain a stalemate. If one side gains an advantage, though, it's likelier to be Russia. They might capture a non-trivial amount of additional land at terrible cost to themselves.

Israel will end its major ground operations in Gaza early next year and switch to a prolonged period of lower-intensity operations.

I don't know who will win the 2024 U.S. Presidential election. The odds are 85% that it will be a Biden-Trump rematch, in which case I give Biden a slight advantage.
There's a 10% chance either Biden or Trump won't make it to the general election, and a 5% chance neither of them will.
In other words, I'm making the somewhat bold prediction that the current conventional wisdom about a Biden-Trump rematch being a certainty is wrong.

Progress with LLMs and other types of narrow AIs will continue at roughly 2023 rates. I don't know if GPT-5 will be released in 2024 and don't think it's that important. For sure, AGI will not be invented by the end of 2024.

2024 is now at an end, so let me rate my predictions:

1) I was right about the Ukraine War, and people of different perspectives could argue that either of the possibilities I described materialized. Russia has gained more Ukrainian territory this year than it did in 2023, but also lost more men and weapons for it. The amount of territory gained may or may not be significant depending on how you look at it, though at this pace of advance, Russian troops won't get to Kiev for decades.

2) I was wrong about the Gaza War: Israel's troop presence in Gaza is still high and there's no timeline for them to leave. They might create permanent military corridors to block weapons smuggling across the Egyptian border and to hinder the free movement of people within Gaza. I didn't understand the mindsets of the two sides and believed that Israel would call it even this year after decimating Hamas and killing 20x as many Palestinians as Israelis who died on October 7.

3) My U.S. Presidential election prediction was mostly right. I made the mistake of giving Biden a small advantage in a matchup with Trump because, like everyone else at the time, I didn't realize how much Biden's cognitive faculties had degraded and how dishonestly the clique of people around him had kept it secret from the American public. I consider the correctness of my prediction that at least one candidate wouldn't make it to the election to be a major success.

4) I was right about LLM progress and GPT-5.

[firestar464]

Joe was just tired. He visibly improved after dropping out.

IIRC the discourse wasn't so much that "everything is fine with him" but rather "it's not unusual for someone at his age and therefore he neither has dementia nor is he unfit to hold office"

[firestar464]

IMO Pekka Haavisto will win the Finnish presidential elections. I'll be back with more predictions

Pakistani elections: Imran Khan's PTI comes out on top (Incorrect due to rigging)
Indonesian elections: Prabowo Subianto wins the presidency as right-wing parties, including Subianto's Gerindra, come out on top in the legislative elections (Correct)
Portuguese elections: The SDP comes out on top (Correct)
Panamanian elections: Former president Ricardo Martinelli is reelected, while his RM party comes out on top in the legislative election (Partially correct; Martinelli was disqualified due to crimes. His ally did win though.)
Lithuanian elections: The LSDP comes out on top (Correct)
Uruguayan elections: The FA wins the presidency while also coming out on top in the legislative elections. (Correct)

[funkervogt]

Joe was just tired. He visibly improved after dropping out.

IIRC the discourse wasn't so much that "everything is fine with him" but rather "it's not unusual for someone at his age and therefore he neither has dementia nor is he unfit to hold office"

Keep telling yourself that!

[Bird]

Joe was just tired. He visibly improved after dropping out.

IIRC the discourse wasn't so much that "everything is fine with him" but rather "it's not unusual for someone at his age and therefore he neither has dementia nor is he unfit to hold office"

I think the fact he ever considered running for re-election at 81 showed a shocking amount of hubris.

Honestly, I'm increasingly leaning towards the opinion that the democrats were fated to lose 2024 regardless of whether they ran Harris, Biden, or someone else. That's just my opinion, though.

Speaking of my opinions, I will follow up shortly with an assessment of my 2024 predictions (TL;DR they weren't as accurate as I'd hoped)

[Bird]

I wish I had more to contribute to these. Most of my predictions are pretty safe/mundane/common anyway, or just focused on stuff I personally care about.


I predict a Biden-Trump rematch in the 2024 election (seems unbelievable typing that, but that's what everything seems to point to right now). Biden will win, but Trump will receive plenty of support despite (I predict) literally being convicted of at least one serious crime. Trump votes will possibly be more than last time (74.2 million). It will probably be a far closer race than a lot of people are expecting. Trump is actually polling BETTER than Biden right now, but I suspect that will change. It absolutely IS possible that Trump wins again, but at this stage I'm predicting otherwise.

If Trump loses, there will be an even more aggressive and far more calculated attempt to overturn the results.


(AMENDMENT, 30 July 2024: Biden has dropped out, stuffing up much of the above prediction. I am predicting a narrow to moderate Kamala Harris victory at this stage. I maintain what I wrote above about it being a fairly close race and Trump making a big effort to overturn the results if he loses.)


Ukraine war will continue past the end of 2024.

GPT-5 will release before the end of the year.

The UK general election itself may not take place this year - but if so, the Conservative party will suffer one of their most significant defeats in history. Unless they pull some seriously impressive politicking and Labour does nothing but f*ck up between then and now, I see no way around that.

Australian Prime Minister Anthony Albanese's support will continue to slide due to his perceived lack of action on housing and cost of living by the Australian public (plus some lingering negative sentiments for the failed Voice campaign). Next election is not until 2025. At this stage I'm predicting sliding support for the "Big Two" parties and increases for minor parties across the political spectrum (though part of this may be just my wishful thinking.)

Amazon's "Fallout" TV show will be positively received by fans.

"Furiosa: A Mad Max Saga" will be well-received also.

f*ck it... I'll predict Grand Theft Auto 6 will release before the end of the year. Plenty of people are predicting 2025, but I'm going to lock in 2024. It will - most likely - be the biggest launch in video game history to date.


I haven't really typed up one of these before. You can bet I'll be back here December 2024 to see how right or wrong I was.

-------------------------------
1. I was INSANELY off the mark about the U.S election. I am currently thinking this is the the worst prediction I've ever made. What I did get right was Trump getting convicted of a serious crime and getting more votes than last time. What I did not fathom was Trump actually winning a second time, complete with a moderate popular vote win by a moderate 1.5%. I'm not American but still feeling a little stunned by this. He won fair and square, he was ultimately what the people wanted. Better the clearly nuts, convicted felon than a mediocre democrat, I guess? Coping and seething aside, if nothing else this is certainly the worst election prediction I have ever made. I got cocky by correctly predicting the US result in 2020, and two other world federal elections in 2022 and 2024. I have to face the fact that I let my bias get in the way (something fellow wrong person Allan Lichtman is determined not to admit). Normally I would step back from election predictions after taking such a massive L, but… well, my country has one in about 5 months. I can't help it.

2. Correct about the Ukraine war (unless it magically resolves in the next ten days). Sadly. I could have accepted being wrong about that.

3. Incorrect about GPT-5, which is slated for early 2025 I think.

4. Correct about the UK election, which was indeed a Labour party win and was (by some metrics) the worst defeat the Conservatives have ever suffered. Notably, the election was also a historic result for small parties and a record low turnout for Britain's big two (Labour & Conservative)

5. Mostly correct about my domestic (Australian) political predictions. Albanese's support has continued to tank (this does not surprise me) and he has now lost the two party preferred vote. Unfortunately, the minor parties don't seem to have had much change in polling. Sigh. Welcome Prime Minister Peter Dutton, the (probable) next leader of this dumbass country.

6. Correct about the Fallout show which was generally well-regarded, barring a few puzzling story choices (No matter how much people try to justify it, I personally think the destruction of a certain lore-critical location was pretty stupid). It won several different awards.

7. Correct about the Furiosa movie. Notably, while reception to the movie was almost universally positive, it bombed at the box office. People didn’t turn out for it. Not personally sure why but I can think of a few reasons.

8. Completely incorrect about Grand Theft Auto 6, and in hilarious fashion. I think we got a trailer with a 2025 release date less than a day or so after I made my initial post. As far as we know the game is slated for late 2025, but some people seem to think a delay to 2026 is likely.


OVERALL ASSESSMENT: Mixed I suppose. Most of the things I got right were fairly predictable and I absolutely bombed the US election predictions. At least I didn't write "Joker: Folie à Deux releases to major critical and commercial success".

I'll drop into the 2025 predictions thread at some point with my contribution.

[firestar464]

Joe was just tired. He visibly improved after dropping out.

IIRC the discourse wasn't so much that "everything is fine with him" but rather "it's not unusual for someone at his age and therefore he neither has dementia nor is he unfit to hold office"

Keep telling yourself that!

ok