Building @circlebackai. Dangerous Optimism. Hacker
San Francisco, CA 🇦🇷
Joined June 2022
- Tweets276
- Following340
- Followers237
- Likes693
fuck your motivation. If you ever worked hard, or wanted to achieve something big, you can tell motivation is a SUPER ephemeral state that you randomly encounter yourself in and that you need hard discipline always
3
220
Was revisiting this lecture today, proactive agents are something that I'm really interested in. Started really thinking about it ~1.5 years ago, when I was doing growvy (teapaw.org/GadiBorovich/status/20…) and wanted this kind of always-on feeling that @bot, dots and many others are trying to achieve today.
I think we are still not close to get it, how can proactive be a "routine" or a "schedule" which is at high level just a cron job that wakes the agent up, I did like the webhooks that grok bot has, same as I thought for growvy, but thats still not enough. Outside coding, the models are still dumb, forget a lot of things when multi tasking/long tasking. I encountered myself a lot of times with my group of bots having to remember THEM about something they should have known.
How do always-on, proactive agents like @Bot work?
Watch my lecture at Stanford CS146S:
2:35 How we got here
6:13 What changed in the models
10:06 Inside an always-on agent
18:07 The harness
27:26 Context engineering
34:26 Where this is going
1
4
376
ponti retweeted
companies are defaulting to recording everything
most of what’s recorded isn’t shared
we shipped folders in Circleback to fix that
Since ~3 weeks ago, I would say that 80% of my workflow goes through grok bot, it's just a super convenient UI and experience. For both work and personal stuff. I didn't try claude duet yet
What's New in Grok Bot
00:14 Shopping cards
00:27 Product pages
00:39 Order tracking
00:53 Payments
01:33 Custom modes
01:52 Team Bot logins and setup
02:09 Share and find Team Bots
02:51 Connector asks in Slack
03:12 Contacts in Vault
03:31 Rapid fire
1
2
574
Been using @eve since v0.11.10 when I built an internal agent for @circlebackai , was really bad. The amount of improvement they did is huge!
2
21
1,655
would be really cool to get Neuralink soon, plus API access so I can log every thought, conversation and neuron firing
Early Neuralink speech restoration data:
We are enabling speech to text and voice at conversational rates in participants with ALS. One participant achieved 75 WPM at 5.8% WER on a 125k-word vocabulary in their first week.
1
109
sam did it again
Announcing Artificial Analysis Intelligence Index v5, featuring 2 new benchmarks and 3 upgrades. For the first time, the majority of the Index is built on private datasets
The Artificial Analysis Intelligence Index is our composite measure of model capabilities across tasks reflecting real-world usage today. v5 is our biggest upgrade since v4 launched in January. With v5, 60% of the Index is now based on evaluations with private test sets or held-out answers.
Alongside Intelligence Index v5, we’re launching the Artificial Analysis Progress Index v1, which tracks model capability over time on one continuous scale.
New to the Intelligence Index:
➤ FrontierCode 2.0: Cognition's newly-launched private benchmark of real-world coding tasks, graded on whether a repository maintainer would merge the agent's change
➤ Terminal-Bench-Science: 70 agentic scientific research tasks, run in a terminal environment
Upgraded:
➤ HLE-Diamond (replacing Humanity's Last Exam): a refined 1,000-question version of Humanity's Last Exam, split evenly between reasoning and knowledge questions
➤ AA-Omniscience v1.1: An updated dataset with saturated and public set questions removed, new questions added, and ambiguous or erroneous questions revised
➤ CritPt v2: a private, 40-question hard subset of CritPt v2, revised by the CritPt authors and run with a code-execution tool
Staying in the Index: AA-Briefcase, AutomationBench-AA, GDP.pdf and Terminal-Bench 4.0
Removed from the Index: GDPval-AA v2.1, SciCode, and AA-LCR v1.1
Key results:
➤ Claude Opus 5.5 (max) leads at 58, ahead of GPT-6 Astra (max, 57) and GPT-6.1 Sol (max, 55): Anthropic and OpenAI hold the top five models. Gemini 4 Argon (high) is next at 50
➤ The open weights gap widens from 3 points to 12: Kimi K3.1 Pro (weights not yet released) is the top open weights model at 46. Among released weights, DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) lead at 32 - 26 points behind the leader and below Claude Haiku 5.5 (max, 36)
➤ Cost per task varies by more than 13x among the leading models: of the 10 highest-scoring models shown, GPT-6.1 Terra (max) is the cheapest at $0.96 per Intelligence Index task and Claude Fable 5.1 (max) the most expensive at $13.04
92
