Ansgar

@ansgar.eth

213 Following

122972 Followers

Ansgar pfp

Haven't found a good way yet to handle cross-posting here and on twitter, so you can find some 3074 nostalgia tweeting here: https://twitter.com/adietrichs/status/1779969081711431883

4 replies

2 recasts

39 reactions

Ansgar pfp

Throwback to when we were accused of attacking the Ethereum governance process for proposing EIP-3074: https://ethereum-magicians.org/t/protecting-the-eip-process-from-special-interests-examples-case-study/6542

2 replies

1 recast

57 reactions

Ansgar pfp

I have mixed feelings about EIP-3074 finally on path to reaching mainnet, almost 4 years after we originally proposed it. We fought so hard for it back in 2021, to no avail. Now that we are much further on the account abstraction roadmap, I am less confident that the focus on improving EOAs is still warranted.

5 replies

2 recasts

50 reactions

Ansgar pfp

New mixtral-8x22b just dropped. And I didn’t even have time to check out Command R+ from last week yet. And (small) Llama 3 apparently coming in a week or so. Looks like an exciting next wave of open-weight LLMs! https://x.com/mistralai/status/1777869263778291896

3 replies

3 recasts

29 reactions

Ansgar pfp

Today’s vibes: On a walk in Rome at 1am listening to a "Why Rome Fell" podcast.

11 replies

3 recasts

89 reactions

Ansgar pfp

Would be really useful to be able to mute individual channels. Follow someone and keep seeing casts you don’t care for in a particular channel? Just mute it!

3 replies

5 recasts

44 reactions

Ansgar pfp

There has been quite a bit of backlash against this recent issuance change proposal by @caspar and me (and maybe rightfully so). I tried to clarify our intention in a likely way too long tweet over on X. Hope that adds some context. https://x.com/adietrichs/status/1775554454923747758

15 replies

17 recasts

120 reactions

Ansgar pfp

I disagree with Charlie here. DAS has received little attention compared to other parts of the roadmap, but there has been a lot of progress over the past year. @fradamt had a great summay post (link below screenshot): https://ethresear.ch/t/from-4844-to-danksharding-a-path-to-scaling-ethereum-da/18046

6 replies

5 recasts

35 reactions

Ansgar pfp

Are you in Europe and are inexplicably not thrilled about mommy EU protecting you from usable search results when you search for anything maps-related? There is a quick fix for Chromium browsers: https://chromewebstore.google.com/detail/edllcgchknhokighleffpipdedmpgiln

1 reply

2 recasts

24 reactions

Ansgar pfp

My take: L2s should target very different full node requirements than L1, as they inherit L1 security. The correct endgame is lightweight L1 nodes and datacenter L2 nodes. Importantly, by running an L1 node, users can still verify proofs about the state of the L2, so the end-to-end user flow remains fully trustless.

10 replies

8 recasts

61 reactions

Ansgar pfp

brought to you by diet coke

5 replies

3 recasts

39 reactions

Ansgar pfp

Thanks to @joncharbonneau and @hasu for having us on the podcast! It really is an important topic, and chatting with you was a pleasure as always! It's 2.5 hours, but I hope some people will find the time to listen! I thought the second half of the episode was especially productive.

1 reply

2 recasts

21 reactions

Ansgar pfp

OH: This space is so ripe for disruption by professionals, it's insane.

8 replies

3 recasts

48 reactions

Ansgar pfp

My hope is that the current (brief) phase of too-cheap-to-meter 4844 DA will serve as a wake-up call, by illustrating the bottlenecks current EVM equivalent L2s will run into once data is no longer scarce. I think we have been too complacent about pushing towards a fundamentally sound high-throughput L2 architecture.

3 replies

3 recasts

24 reactions

Ansgar pfp

There is a fundamental tension here that the cheap blobs surfaced: "EVM equivalent" L2s have similar bottlenecks to L1, but might want to be significantly more ambitious around e.g. throughput. In that case, waiting for L1 to solve these bottlenecks might not work, as L1 is happy at current levels.

1 reply

1 recast

18 reactions

Ansgar pfp

At university, I took a class on "brain computer interface". In the first lecture, the professor showed a short video of a monkey with a chip implanted. He pointed out that invasive BCIs are much more precise, but infeasible for humans. Then we spent the semester dealing with noise filtering for non-invasive human BCI.

0 reply

0 recast

12 reactions

Ansgar pfp

Neuralink just streamed a short video with their first human patient playing chess and controlling his MacBook with his brain only: https://x.com/neuralink/status/1770563939413496146 Of course we are in the very early days of this particular branch of humanity's tech tree, but I can't wait to see where this is going!

2 replies

3 recasts

19 reactions

Ansgar pfp

Strongly agree with pg here. We have one existence proof for an "asshole" Elon, and zero for a "nice guy" version. That should make the null hypothesis that the personality traits that sometimes make him so annoying are also key factors in his extraordinary achievements. https://x.com/paulg/status/1768611682085937215

7 replies

1 recast

17 reactions

Ansgar pfp

On the one hand, I think the current excitement about super cheap L2 transactions is dangerous, as those fee levels are not yet sustainable. On the other hand, it is an amazing illustration of the modular architecture endgame: Once we reach the end of the Danksharding roadmap, <$0.01 transactions will be the norm.

0 reply

2 recasts

37 reactions

Ansgar pfp

"Was Dencun already priced in?" - the gc shares its opinion with a modified wsb copypasta

1 reply

1 recast

14 reactions