Search

AGI is approaching, making deeper study of its technical and social consequences urgent.

Aravind reiterates a 2028 AGI forecast and places AI exceeding specialist humans, which he calls ASI, in the 2040s.

AGI has arrived.

Congratulations to our friends at @OpenAI! I guess this makes Abilene the birthplace of AGI!

Agree with @JensenHuang: we’re entering the AGI era.

For me personally, I do think we're there.

Sooner, given progress is happening faster than I expected.

His modal expectation was full superintelligence within two to three years, while acknowledging it might take 10–20.

His median date for transformative AI was 2038, but his uncertainty extended from a few years to more than a century.

The transition from the first automated AI researchers to an intelligence explosion would likely take years rather than months.

And I actually think we crossed that about 3 months ago.

Levy identifies ChatGPT’s 2022 release as AGI’s arrival, correcting his own earlier claim that the release was in 2023.

I'm calling it. AGI is already here – it's just not evenly distributed yet.

I personally have an ASI median of ~2035

I think it's now. I think we've achieved AGI.

When the book is written, AGI Day will be in today’s past.

I think we crossed the AGI line with the GPT 5.2/Opus 4.5/Gemini 3 generation of models

Tesla will be one of the companies to make AGI and probably the first to make it in humanoid/atom-shaping form

Whether technique novelty is needed for AGI is unknown, but I'd guess ~no, thus human level AGI in 3 years likely

Its raw intelligence is at human level.

After AI automates AI research, the remaining window before uncontrollably powerful superintelligence could be around a year or less.

A country of AI geniuses has roughly even odds within one to three years.

AGI ~2030.

OpenAI built AGI first. In like 2020.

We basically have built AGI, or very close to it.

powerful AI could be as little as 1–2 years away

We might have AI that is smarter than any human by the end of this year

AGI is approaching and will profoundly reshape the economy.

There is a fifty percent chance of AGI by 2030.

My predictions for 2026: Coding and Mathematics AGI

Fleek Futurist reiterates a personal forecast of AGI by 2031, saying this has been their prediction for more than a decade.

AGI may already have passed as a recognizable milestone.

In ten more years, I believe we are almost certain to build superintelligence.

There remains a fifty percent chance of minimal AGI by 2028.

A human-like learner that becomes superhuman could arrive in five to twenty years.

My median AGI arrival estimate is now 2035.

around 2030, lots of uncertainty though

AI able to do any human intellectual task was still decades away or longer, despite rapid practical progress.

Frontier AI could match or exceed human capabilities within a few years, requiring safeguards and public participation.

My estimate of the probability of Grok 5 achieving AGI is now at 10% and rising

It will take about a decade to work through all of those issues.

It seems like AGI won't arrive in a year, but it could easily this decade.

Kat gives a 60% chance of AGI within five to twenty years and 90% within three to sixty years.

I now think @xAI has a chance of reaching AGI with @Grok 5.

AGI could arrive within the next decade.

Wait until you see Grok 5. I think it has a shot at being true AGI.

my bottom-line numbers are somewhat different (median ~EOY 2029)

gpt-4o is AGI, its just not the smartest one

AGI could arrive in five to twenty years.

Developing superintelligence is now in sight.

There is a fifty percent chance of AGI by 2030.

The continuous RL improvement of Grok feels like AGI.

AGI is already here. All the components exist; we just need to stitch them together.

Probably smarter than any single human at anything in <2 years

There was a 25% chance of AI fully automating an AI company within four years, and a 50% chance within eight years.

Humanity is close to building digital superintelligence

we have AGI. our best models pass the turing test, this is AGI.

There is roughly a fifty percent chance of AGI in five to ten years.

Autonomous AI self-improvement could be possible within two years; even substantially shorter timelines were hard to exclude.

AGI just before 2030.

I really don’t know, but I do think they can be credibly seen as a form of “Jagged AGI”

His timeline for AI doing every remote job was about five years earlier than Ege Erdil's 2045 estimate.

A drop-in AI replacement able to do every kind of remote work seemed likely around 2045.

I think it is AGI, seriously.

AGI remains roughly five to ten years away.

I don't think it's unreasonable to speculate that we may have ASI in 2027

Todd says AGI before 2030 is difficult to rule out, while explicitly emphasizing that its arrival is uncertain.

2027 is a plausible scenario for AGI followed by rapid superintelligence.

AGI could match the strongest human specialists within three to five years.

Smith calls Claude 3.7 AGI for autonomous coding work, while emphasizing its tendency to fixate on details and lose sight of goals.

Human-level AI could be here in five to ten years.

I'm pretty sure we'll have cheaper than human AGI before 1 decade passes

agi is already here, but it's jagged and doesn't fit in well with our preconceptions.

the final race to A.G.I. is afoot

Systems that start to point toward AGI are coming into view.

AGI will arrive much sooner than two years.

AI smarter than almost all people at almost everything is most likely in 2026–2027.

We're 1-6 years away from AGI

AI could surpass almost everyone at almost every task within two to three years.

We are now confident we know how to build AGI as we have traditionally understood it.

we will have AGI in 4 years

GPT-4 was AGI to me.

AGI is plausible within five to ten years, with substantial uncertainty.

It was talking about predicting AGI within 10-15 years (which is also roughly my prediction).

Superintelligence may arrive in five to twenty years.

I predict AGI in the next 10 years (65%)

Extrapolating current progress suggests powerful AI in 2026 or 2027, possibly with some delay.

AGI.

Within five years, people may look back on an AGI milestone that has already passed.

AGI by 2029 is conservative.

AGI will be achieved in two to three years.

I think it could come as early as 2026

Human-level AI remains years to decades away.

It is possible that we will have superintelligence in a few thousand days

AGI, matching or surpassing human intelligence generally, will arrive by 2029.

DeepMind remains on track for the twenty-year AGI plan it began in 2010.

And it seems like full blown agentic AGI by 2050 is quite likely

Human-level AGI will arrive by 2029.

In less than ten years, AGI will be realized

Current systems already qualified as AGI if the threshold was average performance across common human tasks.

AGI by 2027 is strikingly plausible.

He did not expect AGI the following year, but considered its arrival in two or three years possible.

AGI is already here and if you used Claude in Websim you’d know.

I think it's probably next year, within two years

GPT3 was agi

MacIver would be extremely surprised if AGI had not arrived within fifty years, and expects some connection to the current generation of chatbots.

Human-level or superhuman AGI is likely by 2029–2030, and could arrive as early as 2027.

AI could pass every human test within five years, if that is the definition of AGI.

Bergman gives a 50% chance of AGI by 2027, while emphasizing uncertainty and presenting several other probability thresholds.

AGI capable of any human intellectual task remained at least decades away.

It is increasingly clear that all roads lead to AGI. Tesla is building an extremely compute-efficient mini AGI for FSD.

AGI could be reached in the next four or five years.

AI could match ordinary humans on a broad set of intelligence tests within five years.

I would say now, three to eight years is my take

I think there's a 50% chance that we have AGI by 2028.

Artificial General Intelligence Is Already Here

AGI will surpass humanity's combined intelligence tenfold within ten years.

We are six, seven, eight years out from something really big and important being publicly visible.

A prototype AGI showing early signs of life by 2030 is possible.

Would never have guessed Gary Marcus was at 35% for AGI within 13 years, that's pretty close to my own views

A model resembling a generally educated human in conversation could be two or three years away.

Human-level AI could arrive in the next two decades, possibly within a few years.

Digital superintelligence could arrive in five or six years.

superintelligence could arrive this decade

My current estimate places a 95% confidence interval for the time horizon of superhuman intelligence at 5 to 20 years.

AI exceeding expert ability in most domains could arrive within ten years.

If AI became highly profitable or substantially accelerated AI research this decade, it was hard to see AGI remaining absent in 2040.

I now predict 5 to 20 years but without much confidence.

Very capable, very general AI could arrive within a few years, perhaps a decade.

there's reason to believe we're years rather than decades from getting there

I think there is a 10% chance we have human level AGI in 5 years.

i have been told that gpt5 is scheduled to complete training this december and that openai expects it to achieve agi.

By historical definitions, the simulcra running on GPT3.5/GPT4 are AGI --goalposts are flying.

Now I think it might be 20 years or less.

an early (yet still incomplete) version of an artificial general intelligence (AGI) system

AGI is here. It's not yet superhuman, but it is general intelligence

Shapira expects a scenario involving superintelligent AGI and loss of human control most likely within a few years, rather than decades.

No AGI within next 100 years, absent some kind of civilisational decline/collapse, is super unlikely.

AGI may arrive before this century ends, but current language models lack several essential capabilities.

Maybe a 60% chance of initial AGI by 2030; about 95% by 2050.

My AGI timelines are short-ish (within two decades with things getting weird soon)

AI should pass a rigorous Turing test by 2029.

A 55–60% chance of AGI showing signs of life by 2030.

Like I think on balance we have AGI in 20 years.

agi is already here

I’d be surprised if we don’t have AGI by then.

Solving these scaling challenges is what will deliver AGI.

AGI is coming in 15 years

AI vastly smarter than humans could arrive in under five years.

near term AGI should be taken as a serious possibility

A few years could separate major economic effects from AI doing everything, even under a comparatively gradual takeoff.

Human-level AI is less than ten years away.

Human-level AI in 2030.

Human-level AI might arrive in five years or later.

AGI could happen any day.

I've been consistent in saying 2029.

Human-level AI within several decades is plausible.

Humanlike intelligence and consciousness remain over a century away.

Computers should equal and surpass humans' remaining advantages by 2029.

Humanlike understanding and reasoning might require another 20–50 years.

Brain-equivalent computers should arrive this century.

General AI could arrive in 10–100 years, with 90% confidence.

Once we have animal-level AI, a few years or decades later we may have human-level AI

We're decades away from anything that's nearing human-level general intelligence.

superhuman machine intelligence (SMI) is probably still decades away

Computers will match human abilities by 2029, including emotion.

By 2029, they will match human intelligence.

I believe that gap will close by 2029.

On that question I've consistently said 2029.

Artificial intelligence will reach human levels by around 2029.

A 50% chance that machines could perform roughly 80% of human jobs by 2050.

Human-level AGI still centered on 2028; an impressive proto-AGI might arrive within eight years.

I'd be surprised if the Technological Singularity hasn't happened by 2030.

A 50% chance of human-level machine intelligence by 2028.

High for the next few decades

if you follow this plan for, let's say, a decade, you'd get there

Roughly 2030.

Kept his human-level AGI forecast at a 2028 mean and 2025 mode.

An adequately funded, concentrated AGI effort might enable a singularity within roughly a decade.

Human-level AGI was most likely around 2025, with an expected date of 2028.

I think we're ‘on track’.

A decade or less could be enough for a positive singularity if society made it a major priority.

By 2029, machines should combine human cognitive flexibility with superior memory and speed.

By 2029 a machine will be able to pass a Turing test

A positive singularity could be achievable within ten years if the right people received sufficient support.

by 2029 computers will be fully equal to humans and thereafter surpass them

by 2029 computers will pass the Turing test

Bet that a machine would pass the agreed Turing test by the end of 2029.

we will achieve machines in a few decades that have the complexity and richness of human thinking

Robots could reach human-level intelligence around 2040.

By 2029, the software needed for intelligence would be largely mastered.

By 2029, computers would routinely pass apparently valid Turing tests, while debate about the full breadth of their intelligence continued.

There was a case for superhuman artificial intelligence within the first third of the twenty-first century.

Robots would reach human-level capabilities by the middle of the next century, in software as well as processing power.

By around 2040, there will be no job that people can do better than robots.

Robots combining world simulation and reasoning could gain most human abilities in the 2030–2040 period.

Technology would create greater-than-human intelligence within the next thirty years.

Established, broadly accepted success on a full Turing test would arrive sometime between 2020 and 2070.

Human-level intelligence would require advances extending well beyond the hardware he expected around 2017.

Human-level problem solving in several technical fields could follow a general AI theory by five to ten years plus a few more.

Superhuman intelligences should be created between twenty and one hundred years from now.

The required breakthroughs could take anywhere from five to five hundred years.

If hardware projections held and software kept pace, machines thinking as well as humans could begin appearing in ten years.

An ultraintelligent machine might arrive around 1994, with ten years of uncertainty in either direction.

Within a generation, artificial intelligence would be substantially solved and few intellectual domains would remain exclusively human.

It was more likely than not that an ultraintelligent machine would be built within the twentieth century.

An exceptional machine intellect capable of redesigning itself might emerge in 1978.

Machines might exceed human general intelligence within a lifetime.

Within twenty years, machines would be technically capable of any work a person could do.

A useful artificial brain could take roughly 13 to 30 years, conditional on it being achieved within the next century.

There are now machines that think, learn and create; their range of problems will expand toward the human range.

An unrestricted imitation-game contest would be at least a century away.

If machines capable of thinking were built, they would probably surpass human intellectual powers soon afterward.

Around 2000, computers could make an average interrogator misidentify them at least 30% of the time in a five-minute imitation game.