Ben Goertzel ✓
@bengoertzel · Vashon Island, WA, USA · joined 02 Apr 2008
How can we best create beneficial AGI, ASI and Singularity for all sentient beings?? -- Founder/CEO @asi_alliance @singularitynet, BGI Foundry, etc. etc.
78 224Followers
1 685Following
8 866Posts total
140.8KViews on collected posts
Ultimi post
Cerchiamo nuovi post…
As part of our beefing up of the AGI Society, we have started publishing our best attempt at objective evaluations of "claims of human-level or stronger AGI." .... While we don't think anyone is there yet, things are getting close enough that making this sort of careful and
7.5K views · 84 likes
· 12 reposts · 7 replies
Open on X →
We just released an evaluation examining the publicly available evidence for the proposition claimed by @JensenHuang and @gdb that GPT-6 Astra has achieved AGI
https://t.co/2KzpN77XQa
8.4K views · 19 likes
· 4 reposts · 0 replies
Open on X →
Thoughts on OpenAI’s Millennium Prize result, formal verification as a key to safe superintelligence, and why coming up with new math ideas is the harder problem. https://t.co/M0N1omR1H6
4K views · 69 likes
· 8 reposts · 9 replies
Open on X →
@bengoertzel "kind of" is doing a lot of work.
the tools got scary-good at tasks. they did not get a self. still a very good intern with no shame.
129 views · 1 likes
· 0 reposts · 0 replies
Open on X →
@bengoertzel The distinction between narrow model performance and true general intelligence remains unproven in practice.
105 views · 1 likes
· 0 reposts · 0 replies
Open on X →
@bengoertzel the term drifted so far from your original coining, now every model launch slaps AGI on it for the headline, not the actual bar
197 views · 1 likes
· 0 reposts · 1 replies
Open on X →
On GPT-6 Astra, the AGIPO, and what the term AGI was coined to mean. https://t.co/WFQm4KmPQF
12.5K views · 118 likes
· 20 reposts · 7 replies
Open on X →
@bengoertzel Treating agentic code generation as a mechanism to incrementally self-uplift beats blindly throwing 25 dense research papers into a context window and hoping a general intelligence magically emerges.
125 views · 1 likes
· 0 reposts · 0 replies
Open on X →
10. On values: we're shifting control from the LLM in the loop to the value-reflective AtomSpace nervous system in each hive, so the metahive blends humanity's teeming diversity of values into a global-brain value system.
Big 3-6-12 months ahead. My human and AI friends: This is
1K views · 12 likes
· 0 reposts · 2 replies
Open on X →
9. And this can happen everywhere at once. Let 10,000 OmegaHives bloom, each self-uplifting -- connected via OmegaBuzz into a metahive that shares lessons & asks itself for advice, running decentralized on SingularityNET today & on ASI Chain at mainnet. An emerging
1.1K views · 11 likes
· 0 reposts · 1 replies
Open on X →
8. The punchline is the virtuous cycle: every mechanism the hive integrates makes it smarter at implementing the next one, & every eval produces cognitive capital. Humans answer questions at stuck-points instead of typing code — the most efficient use of researchers any dev
302 views · 8 likes
· 0 reposts · 1 replies
Open on X →
7. How do you know the tests themselves are any good? Train narrow AIs on each one — ML, Bayesian, evolutionary — as baselines AND as test-tuners: not too easy, not too hard, no shortcuts. Then wrap the qualified specialists as modules the hive itself can learn to use.
304 views · 8 likes
· 0 reposts · 1 replies
Open on X →
6. AGI Maze (a 2D world harder in many ways than ARC-AGI-3), Neoterics (3D synthetic ecology), RoboGarden, RepoOps, LeanGarden (theorem proving), Virtual Scientist, SocietyLab, HiveForge, SelfLab — plus a Transfer Ring measuring how knowledge flows between all of them.
312 views · 11 likes
· 0 reposts · 1 replies
Open on X →
5. Evaluate against what? Not an AGI score -- benchmarks get hacked; that's what they're for these days. You want an evaluation ecology. I tried to design simple, elegant tests and failed ...so with a few LLMs & agents I built a big nasty but very useful conglomeration of 10
342 views · 11 likes
· 0 reposts · 1 replies
Open on X →
4. The fix is a governed, guided improvement loop: baseline your hive => add ONE cognitive mechanism (uncertain reasoning, attention economics, predictive coding...) => evaluate => tune => promote or park => repeat.
Incremental & cumulative — the way complex systems have always
384 views · 11 likes
· 0 reposts · 1 replies
Open on X →
3. What could possibly go wrong? Whellp, yeah about that .. between a component that works in a small test (with nice math, even!) and one that works at scale lies a LOT of fiddling — algorithms, representations, interfaces. Try that blindly for dozens of components at once &
870 views · 12 likes
· 0 reposts · 1 replies
Open on X →
2. At AGI-26 last week I saw a lot of people asking their agent swarms to code AGI for them: download 25 of Ben's (or other cool AGI) papers, bash them together into a codebase, make it tick. This has in it the seeds of a workable approach to AGI.
1.7K views · 17 likes
· 0 reposts · 1 replies
Open on X →
1. Can a proto-AGI help build its successor? New substack post on how our OmegaHive agent hives can incrementally self-uplift toward AGI — implementing, testing & tuning cognitive mechanisms from decades of AGI research, one measured step at a time... . .
96.2K views · 62 likes
· 9 reposts · 9 replies
Open on X →
Rispetto ad account della stessa dimensione
19 post degli ultimi 90 giorni, accanto alla fascia di 10K–100K follower. raggiunge meno persone di account della stessa dimensione.
Visualizzazioni mediane870questo account924mediana per 10K–100K
Copertura, %1.11%questo account3.62%mediana per 10K–100K
Interazione, %1.37%questo account1.52%mediana per 10K–100K
| Metrica | Questo account | Mediana per 10K–100K | Rapporto |
|---|---|---|---|
| Visualizzazioni mediane per post | 870 | 924 | 0.94× |
| Copertura (visualizzazioni ÷ follower) | 1.11% | 3.62% | 0.31× |
| Tasso di interazione | 1.37% | 1.52% | 0.90× |
Altri account di questa fascia → Confronta con un altro account → Come sono costruiti questi parametri →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
3127 Aug
304
302
1.1K
1K
125
5.4K5 Sep
12.5K8 Sep
197
105
129
4K10 Sep
8.4K11 Sep
7.5K
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
3.85%7 Aug
2.96%
2.98%
1.12%
1.37%
0.80%
1.79%5 Sep
1.16%8 Sep
1.02%
0.95%
0.78%
2.16%10 Sep
0.27%11 Sep
1.38%
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.
What the audience does
Likes70.3%532 in total
Reposts8.7%66 in total
Replies6.9%52 in total
Bookmarks14.1%107 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Account simili
Amit Goenka
@AmitGoenka_
156.4K followers
· 2.5K posts
Anna Gabriel Sabaté
@AnnaGaSabate
156.4K followers
· 6.6K posts
Fortnite News ✓
@Guille_GAG
156.4K followers
· 18.1K posts
Intel News
@intelnews
156.4K followers
· 8.5K posts
2.5次元の誘惑_公式（にごリリ）👼
@ririsa_official
156.4K followers
· 4.3K posts
Barceló Hotel Group ✓
@barcelohoteles
156.4K followers
· 17.7K posts
Icona Pop
@iconapop
156.4K followers
· 10.4K posts
AT&T Mx Ayuda
@ATTMxAyuda
156.4K followers
· 438.8K posts