7 ms·

This is hilariously over-optimistic on the timescales. Like on this timeline we'll have a Mars colony in 10 years, immortality drugs in 15 and Half Life 3 in 20

This is hilariously over-optimistic on the timescales. Like on this timeline we'll have a Mars colony in 10 years, immortality drugs in 15 and Half Life 3 in 20.

sva_ 2y ago
You forgot fusion energy
- klabb3 2y ago
  Quantum AI powered by cold fusion and blockchain when?
zvitiate 2y ago
No, sooner lol. We'll have aging cures and brain uploading by late 2028. Dyson Swarms will be "emerging tech".
mchusma 2y ago
I like that the "slowdown" scenario has by 2030 we have a robot economy, cure for aging, brain uploading, and are working on a Dyson Sphere.
- Aurornis 2y ago
  The story is very clearly modeled to follow the exponential curve they show. Like the drew the curve out into the shape they wanted, put some milestones on it, and then went to work imagining what would happen if it continued with a heavy dose of X-risk doomerism to keep it spicy. It conveniently ignores all of the physical constraints around things like manufacturing GPUs and scaling training networks.
  - joshjob42 2y ago
    https://ai-2027.com/research/compute-forecast https://ai-2027.com/research/compute-forecast In section 4 they discuss their projections specifically for model size, the state of inference chips in 2027, etc. It's largely pretty in line with expectations in terms of the capacity, and they only project them using 10k of their latest gen wafer scale inference chips by late 2027, roughly like 1M H100 equivalents. That doesn't seem at all impossible. They also earlier on discuss expectations for growth in efficiency of chips, and for growth in spending, which is only ~10x over the next 2.5 years, not unreasonable in absolute terms at all given the many tens of billions of dollars flooding in. So on the "can we train the AI" front, they mostly are just projecting 2.5 years of the growth in scale we've been seeing. The reason they predict a fairly hard takeoff is they expect that distillation, some algorithmic improvements, and iterated creation of synthetic data, training, and then making more synthetic data will enable significant improvements in efficiency of the underlying models (something still largely in line with developments over the last 2 years). In particular they expect a 10T parameter model in early 2027 to be basically human equivalent, and they expect it to "think" at about the rate humans do, 10 words/second. That would require ~300 teraflops of compute per second to think at that rate, or ~0.1H100e. That means one of their inference chips could potentially run ~1000 copies (or fewer copies faster etc. etc.) and thus they have the capacity for millions of human equivalent researchers (or 100k 40x speed researchers) in early 2027. They further expect distillation of such models etc. to squeeze the necessary size down / more expensive models overseeing much smaller but still good models squeezing the effective amount of compute necessary, down to just 2T parameters and ~60 teraflops each, or 5000 human-equivalents per inference chip, making for up to 50M human-equivalents by late 2027. This is probably the biggest open question and the place where the most criticism seems to me to be warranted. Their hardware timelines are pretty reasonable, but one could easily expect needing 10-100x more compute or even perhaps 1000x than they describe to achieve Nobel-winner AGI or superintelligence.
    - tsurba 2y ago
      I don’t believe so. I think all important parts that each need to be scaled to advance significantly in the LLM paradigm are at or near the end of the steep part of the sigmoid: 1) useful training data available in the internet 2) number of humans creating more training data ”manually” 3) parameter scaling 4) ”easy” algorithmic inventions 5) available+buildable compute ”Just” needing a few more algorithmic inventions to keep the graphs exponential is a cop out. It is already obvious that just scaling parameters and compute is not enough. I personally predict that scaling LLMs for solving all physical tasks (eg cleaning robots) or intellectual pursuits (they suck at multiplication) will not work out. We will get better specialized tools by collecting data from specific, high economic value, constrained tasks, and automating them, but scaling a (multimodal) LLM to solve everything in a single model will not be economically viable. We will get more natural interfaces for many tasks. This is how I think right now as a ML researcher, will be interesting to see how wrong was I in 2 years. EDIT: addition about latest algorithmic advances: - Deepseek style GRPO requires a ladder of scored problems progressively more difficult and appropriate to get useful gradients. For open-ended problems (like most interesting ones are) we have no ladders for, and it doesn’t work. In particular, learning to generate code for leetcode problems with a good number of well made unit tests is what it is good for. - Test-time inference is just adding an insane amount of more compute after training to brute-force double-check the sanity of answers Neither will keep the graphs exponential.
- km144 2y ago
  The true absurdity of this timeline is that everyone is behaving as perfectly rational players, which is absurd when you look at the current presidential administration, particularly in light of recent events.
ctoth 2y ago
Can you share your detailed projection of what you expect the future to look like so I can compare?
- Gud 2y ago
  Slightly slower web frameworks by 2026. By 2030, a lot slower.
- IshKebab 2y ago
  Sure 5 years: AI coding assistants are a lot better than they are now, but still can't actually replace junior engineers (at least ones that aren't shit). AI fraud is rampant, with faked audio commonplace. Some companies try replacing call centres with AI, but it doesn't really work and everyone hates it. Tesla's robotaxi won't be available, but Waymo will be in most major US cities. 10 years: AI assistants are now useful enough that you can use them in the ways that Apple and Google really wanted you to use Siri/Google Assistant 5 years ago. "What have I got scheduled for today?" will give useful results, and you'll be able to have a natural conversation and take actions that you trust ("cancel my 10am meeting; tell them I'm sick"). AI coding assistants are now very good and everyone will use them. Junior devs will still exist. Vibe coding will actually work. Most AI Startups will have gone bust, leaving only a few players. Art-based AI will be very popular and artists will use it all the time. It will be part of their normal workflow. Waymo will become available in Europe. Some receptionists and PAs have been replaced by AI. 15 years: AI researchers finally discover how to do on-line learning. Humanoid robots are robust and smart enough to survive in the real world and start to be deployed in controlled environments (e.g. factories) doing simple tasks. Driverless cars are "normal" but not owned by individuals and driverful cars are still way more common. Small light computers become fast enough that autonomous slaughter it's become reality (i.e. drones that can do their own navigation and face recognition etc.) 20 years: Valve confirms no Half Life 3.
  - archagon 2y ago
    > Small light computers become fast enough that autonomous slaughter it's become reality This is the real scary bit. I'm not convinced that AI will ever be good enough to think independently and create novel things without some serious human supervision, but none of that matters when applied to machines that are destructive by design and already have expectations of collateral damage. Slaughterbots are going to be the new WMDs — and corporations are salivating at the prospect of being first movers. https://www.youtube.com/watch?v=UiiqiaUBAL8 https://www.youtube.com/watch?v=UiiqiaUBAL8
Trumpion 2y ago
We currently don't see any ceiling if this continues in this speed, we will have cheaper, faster and better models every quarter. Therewas never something progressing so fast It would be very ignorant not to keep a very close eye on it There is still a a chance that it will happen a lot slower and the progression will be slow enough that we adjust in time. But besides AI we also now get robots. The impact for a lot of people will be very real
turnsout 2y ago
IMO they haven't even predicted mid-2025. > Coding AIs increasingly look like autonomous agents rather than mere assistants: taking instructions via Slack or Teams and making substantial code changes on their own, sometimes saving hours or even days. Yeah, we are so not there yet.
- Tossrock 2y ago
  That is literally the pitch line for Devin. I recently spoke to the CTO of a small healthtech startup and he was very pro-Devin for small fixes and PRs, and thought he was getting his money worth. Claude Code is a little clunkier but gives better results, and it wouldn't take much effort to hook it up to a Slack interface.
  - turnsout 2y ago
    Yeah, I get that there are startups trying to do it. But I work with Cursor quite a bit… there is no way I would trust an LLM code agent to take high-level direction and issue a PR on anything but the most trivial bug fix.
    - baq 2y ago
      Last year they couldn’t even do a simple fix (they could add a null coalescing operator or an early return which didn’t make sense, that’s about it). Now I’m getting hundreds of LOC of functionality with multiple kLOC of tests out of the agent mode. No way it gets in without a few iterations, but it’s sooo much better than last April.
danpalmer 2y ago
These timelines always assume that things progress as quickly as they can be conceived of, likely because these timelines come from "Ideas Guys" whose involvement typically ends at that point. Orbital mechanics begs to disagree about a Mars colony in 10 years. Drug discovery has many steps that take time, even just the trials will take 5 years, let alone actually finding the drugs.
- wkat4242 2y ago
  Didn't the covid significantly reduce trial times? I thought that was such a success that they continued on the same foot.
  - pama 2y ago
    No it didn’t. At least not for new small molecule drugs. It did reduce times a bit for the first vaccines because there were many volunteers available, and it did allow some antibody drug candidates to be used before full testing was complete. The only approved small molecule drug for covid is paxlovid, with both components of its formulation tested on humans for the first time many years before covid. All the rest of the small molecule drugs are still in early parts of the pipeline or have been abandoned.
  - danpalmer 2y ago
    The other reply has better info on covid specifically, but also consider that this refers to "immortality drugs". How long do we have to test those to conclude that they do in fact provide "immortality"? Now sure, they don't actually mean immortality, and we don't need to test forever to conclude they extend life, but we probably do have to test for years to get good data on whether a generic life extension drug is effective, because you're testing against illness, old age, etc, things that take literally decades to kill. That's not to mention that any drug like that will be met with intense skepticism and likely need to overcome far more scrutiny than normal (rather than the potentially less scrutiny that covid drugs might have managed).
    - sejje 2y ago
weregiraffe 2y ago
And Winds of Winter when?

This is hilariously over-optimistic on the timescales. Like on this timeline we'll have a Mars colony in 10 years, immortality drugs in 15 and Half Life 3 in 20.

sva_ 2y ago

You forgot fusion energy

klabb3 2y ago
Quantum AI powered by cold fusion and blockchain when?

zvitiate 2y ago

No, sooner lol. We'll have aging cures and brain uploading by late 2028. Dyson Swarms will be "emerging tech".

mchusma 2y ago

I like that the "slowdown" scenario has by 2030 we have a robot economy, cure for aging, brain uploading, and are working on a Dyson Sphere.

Aurornis 2y ago
The story is very clearly modeled to follow the exponential curve they show. Like the drew the curve out into the shape they wanted, put some milestones on it, and then went to work imagining what would happen if it continued with a heavy dose of X-risk doomerism to keep it spicy. It conveniently ignores all of the physical constraints around things like manufacturing GPUs and scaling training networks.
- joshjob42 2y ago
  https://ai-2027.com/research/compute-forecast https://ai-2027.com/research/compute-forecast In section 4 they discuss their projections specifically for model size, the state of inference chips in 2027, etc. It's largely pretty in line with expectations in terms of the capacity, and they only project them using 10k of their latest gen wafer scale inference chips by late 2027, roughly like 1M H100 equivalents. That doesn't seem at all impossible. They also earlier on discuss expectations for growth in efficiency of chips, and for growth in spending, which is only ~10x over the next 2.5 years, not unreasonable in absolute terms at all given the many tens of billions of dollars flooding in. So on the "can we train the AI" front, they mostly are just projecting 2.5 years of the growth in scale we've been seeing. The reason they predict a fairly hard takeoff is they expect that distillation, some algorithmic improvements, and iterated creation of synthetic data, training, and then making more synthetic data will enable significant improvements in efficiency of the underlying models (something still largely in line with developments over the last 2 years). In particular they expect a 10T parameter model in early 2027 to be basically human equivalent, and they expect it to "think" at about the rate humans do, 10 words/second. That would require ~300 teraflops of compute per second to think at that rate, or ~0.1H100e. That means one of their inference chips could potentially run ~1000 copies (or fewer copies faster etc. etc.) and thus they have the capacity for millions of human equivalent researchers (or 100k 40x speed researchers) in early 2027. They further expect distillation of such models etc. to squeeze the necessary size down / more expensive models overseeing much smaller but still good models squeezing the effective amount of compute necessary, down to just 2T parameters and ~60 teraflops each, or 5000 human-equivalents per inference chip, making for up to 50M human-equivalents by late 2027. This is probably the biggest open question and the place where the most criticism seems to me to be warranted. Their hardware timelines are pretty reasonable, but one could easily expect needing 10-100x more compute or even perhaps 1000x than they describe to achieve Nobel-winner AGI or superintelligence.
  - tsurba 2y ago
    I don’t believe so. I think all important parts that each need to be scaled to advance significantly in the LLM paradigm are at or near the end of the steep part of the sigmoid: 1) useful training data available in the internet 2) number of humans creating more training data ”manually” 3) parameter scaling 4) ”easy” algorithmic inventions 5) available+buildable compute ”Just” needing a few more algorithmic inventions to keep the graphs exponential is a cop out. It is already obvious that just scaling parameters and compute is not enough. I personally predict that scaling LLMs for solving all physical tasks (eg cleaning robots) or intellectual pursuits (they suck at multiplication) will not work out. We will get better specialized tools by collecting data from specific, high economic value, constrained tasks, and automating them, but scaling a (multimodal) LLM to solve everything in a single model will not be economically viable. We will get more natural interfaces for many tasks. This is how I think right now as a ML researcher, will be interesting to see how wrong was I in 2 years. EDIT: addition about latest algorithmic advances: - Deepseek style GRPO requires a ladder of scored problems progressively more difficult and appropriate to get useful gradients. For open-ended problems (like most interesting ones are) we have no ladders for, and it doesn’t work. In particular, learning to generate code for leetcode problems with a good number of well made unit tests is what it is good for. - Test-time inference is just adding an insane amount of more compute after training to brute-force double-check the sanity of answers Neither will keep the graphs exponential.
km144 2y ago
The true absurdity of this timeline is that everyone is behaving as perfectly rational players, which is absurd when you look at the current presidential administration, particularly in light of recent events.

ctoth 2y ago

Can you share your detailed projection of what you expect the future to look like so I can compare?

Gud 2y ago
Slightly slower web frameworks by 2026. By 2030, a lot slower.
IshKebab 2y ago
Sure 5 years: AI coding assistants are a lot better than they are now, but still can't actually replace junior engineers (at least ones that aren't shit). AI fraud is rampant, with faked audio commonplace. Some companies try replacing call centres with AI, but it doesn't really work and everyone hates it. Tesla's robotaxi won't be available, but Waymo will be in most major US cities. 10 years: AI assistants are now useful enough that you can use them in the ways that Apple and Google really wanted you to use Siri/Google Assistant 5 years ago. "What have I got scheduled for today?" will give useful results, and you'll be able to have a natural conversation and take actions that you trust ("cancel my 10am meeting; tell them I'm sick"). AI coding assistants are now very good and everyone will use them. Junior devs will still exist. Vibe coding will actually work. Most AI Startups will have gone bust, leaving only a few players. Art-based AI will be very popular and artists will use it all the time. It will be part of their normal workflow. Waymo will become available in Europe. Some receptionists and PAs have been replaced by AI. 15 years: AI researchers finally discover how to do on-line learning. Humanoid robots are robust and smart enough to survive in the real world and start to be deployed in controlled environments (e.g. factories) doing simple tasks. Driverless cars are "normal" but not owned by individuals and driverful cars are still way more common. Small light computers become fast enough that autonomous slaughter it's become reality (i.e. drones that can do their own navigation and face recognition etc.) 20 years: Valve confirms no Half Life 3.
- archagon 2y ago
  > Small light computers become fast enough that autonomous slaughter it's become reality This is the real scary bit. I'm not convinced that AI will ever be good enough to think independently and create novel things without some serious human supervision, but none of that matters when applied to machines that are destructive by design and already have expectations of collateral damage. Slaughterbots are going to be the new WMDs — and corporations are salivating at the prospect of being first movers. https://www.youtube.com/watch?v=UiiqiaUBAL8 https://www.youtube.com/watch?v=UiiqiaUBAL8

Trumpion 2y ago

We currently don't see any ceiling if this continues in this speed, we will have cheaper, faster and better models every quarter. Therewas never something progressing so fast It would be very ignorant not to keep a very close eye on it There is still a a chance that it will happen a lot slower and the progression will be slow enough that we adjust in time. But besides AI we also now get robots. The impact for a lot of people will be very real

turnsout 2y ago

IMO they haven't even predicted mid-2025. > Coding AIs increasingly look like autonomous agents rather than mere assistants: taking instructions via Slack or Teams and making substantial code changes on their own, sometimes saving hours or even days. Yeah, we are so not there yet.

Tossrock 2y ago
That is literally the pitch line for Devin. I recently spoke to the CTO of a small healthtech startup and he was very pro-Devin for small fixes and PRs, and thought he was getting his money worth. Claude Code is a little clunkier but gives better results, and it wouldn't take much effort to hook it up to a Slack interface.
- turnsout 2y ago
  Yeah, I get that there are startups trying to do it. But I work with Cursor quite a bit… there is no way I would trust an LLM code agent to take high-level direction and issue a PR on anything but the most trivial bug fix.
  - baq 2y ago
    Last year they couldn’t even do a simple fix (they could add a null coalescing operator or an early return which didn’t make sense, that’s about it). Now I’m getting hundreds of LOC of functionality with multiple kLOC of tests out of the agent mode. No way it gets in without a few iterations, but it’s sooo much better than last April.

danpalmer 2y ago

These timelines always assume that things progress as quickly as they can be conceived of, likely because these timelines come from "Ideas Guys" whose involvement typically ends at that point. Orbital mechanics begs to disagree about a Mars colony in 10 years. Drug discovery has many steps that take time, even just the trials will take 5 years, let alone actually finding the drugs.

wkat4242 2y ago
Didn't the covid significantly reduce trial times? I thought that was such a success that they continued on the same foot.
- pama 2y ago
  No it didn’t. At least not for new small molecule drugs. It did reduce times a bit for the first vaccines because there were many volunteers available, and it did allow some antibody drug candidates to be used before full testing was complete. The only approved small molecule drug for covid is paxlovid, with both components of its formulation tested on humans for the first time many years before covid. All the rest of the small molecule drugs are still in early parts of the pipeline or have been abandoned.
- danpalmer 2y ago
  The other reply has better info on covid specifically, but also consider that this refers to "immortality drugs". How long do we have to test those to conclude that they do in fact provide "immortality"? Now sure, they don't actually mean immortality, and we don't need to test forever to conclude they extend life, but we probably do have to test for years to get good data on whether a generic life extension drug is effective, because you're testing against illness, old age, etc, things that take literally decades to kill. That's not to mention that any drug like that will be met with intense skepticism and likely need to overcome far more scrutiny than normal (rather than the potentially less scrutiny that covid drugs might have managed).
  - sejje 2y ago

weregiraffe 2y ago

And Winds of Winter when?