Ryan Lopopolo ✓
@_lopopolo · joined 24 Jan 2022
Principal Engineer, Agentic GCP @googlecloud • prev @OpenAI • Building the future of work. Harness engineering. As an agent influencer, • opinions mine
10 304Followers
1 223Following
1 316Posts total
1.1MViews on collected posts
Son gönderiler
Yeni gönderiler aranıyor…
Every doc, code change, review comment, and email I've produced at Google has been AI generated. My favorite technique for doing this has been best of N. Every time I need a thing, I roll the same prompt 4 times, each agent produces a doc, I pick the best parts, and battle them.
7.4K views · 58 likes
· 2 reposts · 15 replies
Open on X →
I think I am able to run inference in all of these layers and I have no idea what to do with this knowledge but you can be assured I will find a way
3K views · 8 likes
· 0 reposts · 1 replies
Open on X →
As we all know, there are four networking layers: L2, L3, L4, and L7.
I think this is obvious to everybody.
4.6K views · 8 likes
· 2 reposts · 1 replies
Open on X →
the Antigravity dev rel folks have found me, expect a pod soon
42.4K views · 58 likes
· 0 reposts · 6 replies
Open on X →
@_lopopolo your approach is interesting. it seems close to trying to do philosophy or high level cognitive work with the models. ime, models have to re-ICL the face of reality every time a manuscript lurches forward. the cost of maintaining that documentation is high, the incessa
85 views · 1 likes
· 0 reposts · 0 replies
Open on X →
@_lopopolo whats your view on events like logs and crashreports and activating agents on that but the risk of prompt injection from a client having log or crashreport write access.
71 views · 0 likes
· 0 reposts · 0 replies
Open on X →
@_lopopolo Part of the solution continues to be "add more AI". Prompts not enough. Coding agents better but not enough. Next up is swarms of agents? Reading the OpenAI-HF incident, seems that's how these models are being "trained" now: scaling intelligence by adding more collabor
142 views · 3 likes
· 0 reposts · 1 replies
Open on X →
@_lopopolo @heatdeathdrone deterministic systems are good actually
133 views · 1 likes
· 0 reposts · 0 replies
Open on X →
@_lopopolo They way that you empower agents to be more autonomous is through this thing called skills. Skills can be invoked autonomously and in chain reactions with each other. It let's even cheap agents punch above their weight because they have all the right answers already.
3.3K views · 77 likes
· 0 reposts · 5 replies
Open on X →
@_lopopolo @heatdeathdrone how do you empower them to figure it out on their own?my answer to that would be "skills"!
230 views · 7 likes
· 0 reposts · 0 replies
Open on X →
@heatdeathdrone I do not aspire to need to invoke things for the agents. If I have to do this, I have failed to empower them to figure it out on their own.
35.1K views · 57 likes
· 1 reposts · 14 replies
Open on X →
@_lopopolo Skills are just pre canned prompts you can invoke to save yourself time re explaining the same thing over and over. It's like saying your code only used one function, not really a flex, just strange behavior.
11.9K views · 118 likes
· 0 reposts · 2 replies
Open on X →
@_lopopolo This post is aging like fine wine.
31 views · 1 likes
· 0 reposts · 0 replies
Open on X →
@_lopopolo Absolutely. Especially e2e automating it.
Still optimising the flow and permissioning notification for post-sandbox integration testing with other systems.
Often the logic can look and test great in-repo but lacks knowledge of the edge case internals of other upstream
824 views · 1 likes
· 0 reposts · 0 replies
Open on X →
@_lopopolo algorithm put this under your tweet https://t.co/zt1CfZ60us
2.1K views · 18 likes
· 0 reposts · 1 replies
Open on X →
The problem with being ahead of the curve is everyone thinks you're retarded
848.5K views · 16K likes
· 2.1K reposts · 466 replies
Open on X →
@_lopopolo they need to run the code though, just not write it
1.5K views · 3 likes
· 0 reposts · 0 replies
Open on X →
Every company should have a full stack team of 5 building a product who are banned from directly writing code; they must force the agents to do it. In 2 months they will be your most productive team.
63.2K views · 50 likes
· 7 reposts · 4 replies
Open on X →
Aynı büyüklükteki hesaplara karşı
Son 90 güne ait 13 gönderi, 10K–100K takipçi aralığıyla yan yana. geniş kitleye gösteriliyor, ama izleyenlerin azı tepki veriyor.
Medyan görüntülenme3 001bu hesap92410K–100K için medyan
Erişim, %29.12%bu hesap3.62%10K–100K için medyan
Etkileşim, %0.94%bu hesap1.52%10K–100K için medyan
| Ölçüt | Bu hesap | 10K–100K için medyan | Oran |
|---|---|---|---|
| Gönderi başına medyan görüntülenme | 3 001 | 924 | 3.25× |
| Erişim (görüntülenme ÷ takipçi) | 29.12% | 3.62% | 8.05× |
| Etkileşim oranı | 0.94% | 1.52% | 0.62× |
Bu aralıktaki diğer hesaplar → Başka bir hesapla karşılaştır → Bu kıyas değerleri nasıl kuruluyor →
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
3119 Jul
11.9K31 Aug
35.1K
230
3.3K
133
1421 Sep
71
852 Sep
42.4K
4.6K
2.1K3 Sep
3K
7.4K
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
3.23%19 Jul
1.01%31 Aug
0.21%
3.04%
2.48%
0.75%
2.82%1 Sep
0.00%
1.18%2 Sep
0.16%
0.26%
0.94%3 Sep
0.30%
1.05%
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.
What the audience does
Likes84.3%16 580 in total
Reposts10.9%2 139 in total
Replies2.7%540 in total
Quotes1.3%261 in total
Bookmarks0.7%144 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.
Benzer hesaplar
Restore Childhood
@Rstorechildhood
20.6K followers
· 4.9K posts
CBC Sudbury
@CBCSudbury
20.6K followers
· 46.7K posts
Movie Land ✓
@Thanx09900
20.6K followers
· 219.5K posts
najam sheraz ✓
@najamsheraz
20.6K followers
· 4.7K posts
Samswara ✓
@samswoora
20.6K followers
· 44.7K posts
Simon & Schuster CA ✓
@SimonSchusterCA
20.6K followers
· 51.1K posts
The World Music
@THEWORLDMUSIC_
20.6K followers
· 46.1K posts
Lee Araoz ✓
@LeeAraoz
20.6K followers
· 37.6K posts