Michael
3.3K posts
@0xsseb @0xCircusLover @Sophon I bought a node at pre-sale. How can i do anything with it? Never got anything
English
This couldn’t be farther from the truth.
Yes we let people go - their roles weren’t necessary or performance wasn’t meeting expectations. In addition we've recognized our gaps & have hired a pretty world class consumer team, which everyone will hear more about soon too.
And yes we’ve been quiet (too quiet I admit) as we’ve been very heads down building some new products.
I realise for investors it’s scary and feels like a rug, but this is not the case. Just been building quietly, that’s simply the truth.
In a few weeks you can expect to hear from us and discover Sophon 2.0 and new products we’ve built.
All I can say for now is that I think general blockchain ecosystems are cooked and all of them struggle to find pmf, so last year in Q4 we decided to take matters into our own hands and build consumer products ourselves we think are useful (hence the team headcount reduction).
Products are coming, top talent has joined, Sophon will be reborn.
English
BREAKING: @Sophon laid off half the team, currently slow rugging millions of money, token is down bad, and the team is on their 5th pivot.
here's what the founder said in 2025 vs now.
English
Michael retweetledi
I am so excited to share our new paper in @Nature: the first programmable, site-specific integration of a large DNA payload into T cells in vivo.
A single IV injection results in therapeutic levels of TRAC-targeted CAR T cells in multiple models.
#Ack1" target="_blank" rel="nofollow noopener">nature.com/articles/s4158…
a 🧵
English
Ultimately, digging up teenaged Rama Duwaji’s tweets is not going to turn any of the mayor’s supporters against him.
It is not in pursuit of any greater truth or serious mission.
It’s just trolling, which is pretty much all the pro-Israel journalists have left at this point.
English
@nateliason Genuine question. If i start taking your build apps class now, do you think ill be in a good spot or is that already dated ha
English
Been heads down the last few rebuilding the chat interface of my new content tool from scratch.
The new version is, without exaggeration, an order of magnitude better than last week's.
Ty for your feedback everyone who tried it out last week.
New version is live now!
English
@nateliason @conductor_build Would you add this to your class over cursor?
English
I did not realize how controversial this would be!
If you're still using Claude / Codex in the terminal, you are missing out.
You should absolutely be in @conductor_build
Dan Shipper 📧@danshipper
“The CLI is the Stone Age from two months ago. GUIs are back.” You heard it here first. @nateliason live from @every's Vibe Code Camp:
English
@nateliason Is there anything from your class that you would add today that wasn’t included in the update but that you think is critical?
English
Yeah this was 1,000% worth it.
Separate Claude subscription + Clawd, managing Claude Code / Codex sessions I can kick off anywhere, autonomously running tests on my app and capturing errors through a sentry webhook then resolving them and opening PRs...
The future is here.
Nat Eliason@nateliason
Hired my first employee today
English
One of the best parts about this workflow is that it solves doing high quality work from your phone.
1. Start your work in Claude Code mobile
2. Open a PR
3. Claude does the rest autonomously on GitHub
Nat Eliason@nateliason
I'm excited to share a Claude Code enhancement I've been working on. It dramatically expands Claude's autonomy once you've created a PR, accelerating your work as a solo dev or small team without getting reckless. Problem: The built-in Claude Code Github PR Review leaves a ton of feedback which you then have to comb through, decide what to address, and feed your priorities back into Claude Code. Solution: dangerously-skip-review gives Claude permission to analyze its own feedback, implement the necessary fixes, and create follow up issues to address after the PR is completed. With dangerously-skip-review, here's what happens once you open a PR: 1. Claude will review the PR in detail. 2. Claude will evaluate that review to identify must-fix items and issues that can be addressed later. 3. Claude will solve the must-fix items. 4. Claude will create follow-up issues to be addressed after the PR is merged. 5. Once the PR is merged, Claude will evaluate the follow up issues, decide if they're worth implementing, and if they are, code them up and open new PRs. If they aren't, Claude will close them with an explanation. 6. The cycle repeats for the new PRs. WARNING: This might eat through your usage limits depending on your repository and plan level, and obviously it could break things. Use it responsibly. If you want to try it out, install the Claude Code Github Actions for your repo (/install-github-app in claude code) then run: npx dangerously-skip-review Let me know what you think! PRs welcome for any improvements.
English
@kazuki_nozawa @Annals_Oncology This is really interesting! So if I'm understanding correctly, if you have had progression on T-DXd, but some initial response, this may be an option?
English
This combination might be an option for post-T-DXd treatment in HER2+ MBC.
In HER2CLIMB-02, adding tucatinib to T-DM1 extended PFS (9.5 vs 7.4 months; HR 0.76) and showed activity even in patients with brain metastases, with manageable toxicity. @Annals_Oncology
#fig3" target="_blank" rel="nofollow noopener">annalsofoncology.org/article/S0923-…
English
@Dr_Oncologista In this case it's mesonephric-like adenocarcinoma, initial response with T-DXd, but keeping an eye on what might be other treatment ideas. Appreciate what you do!
English
@Foxhound131 HER2-low MBC: Use HER2-directed ADCs (e.g., T-DXd). For isolated progression, SBRT + continue therapy if stable. Endometrial cancer is different; HER2-targeting mainly in HER2+ breast or serous endometrial tumors.
English
HER2+ MBC patient on 2L T-DXd with 14-mo excellent response. Now oligoprogression in a single lung lesion, other sites stable, treatment tolerated well.
Next step?"
English
@Annals_Oncology @FAndreMD @JavierCortesMD @curijoey Any studies on thing that improve efficacy for HER-2 Low expression?
English
🆕article in press: Pooled analysis by best response to trastuzumab deruxtecan and related biomarkers in patients with HER2-positive metastatic breast cancer from DESTINY-Breast01, Breast02, and Breast03 @FAndreMD @JavierCortesMD @curijoey. shorturl.at/7oWgJ
English
@Dramanyhassany1 What did you uncover and discuss? can you open your DMs?
English
GFH375: KRAS G12Di targets both ‘on’ & ‘off’ states #PDAC #ESMO2025 @OncoAlert
➡️68% >|= 2 lines tx
➡️ORR 41%; mPFS 5.6mths
➡️mOS NR; most AEs grade 1/2
➡️ctDNA - high rates KRAS(not~90%), low tumour suppressor rates
English
@anthonystaj Seems like there is a mixed perspective on this though. Where do you stand?
English
