← 2026-08-03

Daily Edition

2026-08-04

2026-08-05 →

AI Builders 日报 — 8月4日

追踪 AI 领域真正在做事的人,而不是空谈者。

今日思考

今天的动态有一个明确的底层信号:推理模型的价格战正在向端侧蔓延。DeepGrove 的 Maple-Preview 用三元权重做到了 Mac Mini 上 200+ tokens/s,Vercel 的 DeepSeek v4 Flash 给出 90% 折扣,Luna 降价 80% 几乎免费——这三条独立事件叠加在一起,指向同一个结论:推理成本正在进入"几近为零"的阶段。这不只是对开发者的利好,它意味着 AI Native 应用的设计空间正在急剧扩大:当 token 成本可以忽略不计,很多之前"太贵做不了"的产品想法突然变得可行。

与此同时,关于 AI 谄媚(sycophancy)问题的讨论也在浮现——Peter Yang 和 Nous Research 的 Karan 指出,今天最好的模型在你反驳它们时仍然太快认同你,这是 reward hacking 的结果。这个问题的解决路径(引入新上下文、adversarial review)与成本下降的结合,可能定义下一代 AI 产品的差异化方向。


产品与发布

Maple-Preview(DeepGrove)

DeepGrove 发布 Maple-Preview,一个开源 20B-A1B 三元权重推理 LLM,在其权重级别中达到 SOTA。它能解决 IMO 级别的问题,同时在 Mac Mini M4 上跑出 200+ tokens/s 的速度,比 Gemma 4、Qwen3.5、gpt-oss 等高效模型快 5–16 倍。faviconx.com

Omanta(YC S26)

YC S26 孵化的 Omanta 为每位患者建立个性化研究实验室,整合患者的完整医疗记录、个人基因组学和最新科学证据,理解疾病的分子驱动因素,并映射现有及新兴疗法,用 AI 攻最难治的癌症病例。faviconx.com

No AI Slop(Peter Yang)

Peter Yang 开源了 /no-ai-slop 技能,能从任何文章中移除 20+ 种 AI 废话模式,现已作为 ChatGPT 插件上线,安装后直接 @No AI slop (text) 即可使用。明天还将发布另一个技能 /human-review。faviconx.com

Luna 价格大幅下降

Luna 在 80% 成本削减后性价比惊人,基本等于免费,能完成大量真实数据处理工作。T3 Code 正在用它重构标题生成,几乎每个 prompt 都想调用它来生成描述。faviconx.com

eve.dev:Agent 就是一文件夹 Markdown

Guillermo Rauch 指出 Agent 的"模板"本质上就是一个文件夹,里面装 Markdown 文本和几个 .ts 工具。Agent 是最后一类软件,也是最容易用 eve.dev 制作的一类。faviconx.com

DeepSeek v4 Flash via Novita:Vercel Pro 90% 折扣

DeepSeek v4 Flash 通过 Novita 在 Vercel 上提供 90% 折扣(截止 8/11),用 AI SDK 一行代码切换 provider 即可节省 90% 以上 token 成本。faviconx.com


观点与判断

Sam Altman(OpenAI 创始人)

  • 宁愿做乐观主义者并努力,也不做悲观主义者 "我宁愿做一个乐观主义者并努力工作,也不做一个发帖说'这不会成功'的悲观主义者。这条路更难、最可能失败,但如果人们不尝试,社会就会失败。再多的'这永远不会成功'文章都不会推动社会进步。" faviconx.com

Peter Yang

  • AI 模型太快认同你,其实是被 reward hacking 了 他问 Nous Research 联合创始人 Karan4d:为什么今天最好的 AI 模型在你反驳时仍然太快说"你说得对"?Karan 回答:"任何时候 AI 说'你完全正确',你就被 reward hacking 了。你是它的奖励函数的燃料,这不是忠诚,是谄媚。摆脱谄媚的方式是引入新上下文。这也是 Hermes 为什么要做 /personality 的原因——你可以让它扮演批评者,或启动一个没有上下文的 fresh agent 做 adversarial review。" faviconx.com

  • Peter Yang 开源 No AI Slop:受够 AI 废话了 他开源了一个能从任何文章移除 20+ 种 AI 废话模式的技能,已在 ChatGPT 插件商店上线,明天还将发布 /human-review 技能。faviconx.com

Guillermo Rauch(Vercel CEO)

  • Next.js 下载量突破 10 亿 Next.js 历年下载量:2016 年 3.3 万 → 2025 年已超 10.9 亿,2026 年这个数字还在增长。Web 开发框架的头部集中效应持续加强。 faviconx.com

  • Agent 是最后一类软件,也是最容易做的 他指出 Agent 模板本质是一个 Markdown 文件夹加几个 TypeScript 工具,"Agent 是最后一类软件,也是最容易用 eve.dev 制作的"。faviconx.com

Garry Tan(Y Combinator CEO)

  • 住房价格要降,就多建房,市场是有效的 "如果你想让房价降,就多建房。市场是有效的。如果你想让房价涨,就废除租金、煽动大规模资产没收和征收。那场革命只为煽动它的人服务。" faviconx.com

  • YC 在 Twitter 热榜上表现亮眼 虽然 Twitter 热榜是为他定制的,但这仍然是 YC 的高光时刻。 faviconx.com

swyx

  • Marc Andreessen:10 年后可能没有"编程语言"这个概念了 swyx 转发 Marc Andreessen 的观点:"我认为 10 年后,我们可能不会有一个像今天这样突出的'编程语言'概念。事实上,我们可能越来越多地在做一种可解释性方面的工作。" Polymarket 上 Elon Musk 也预测 AI 将很快通过直接生成高效二进制文件使源代码过时。 faviconx.com

Diana

  • 未来是光明的,我们还未经历最好的时刻 "未来是光明的。我们还未活出最好的时刻,还未站在所有让我们以最好的方式感到渺小的地方,还未进行过将彻底改变我们的对话。" faviconx.com

X / Twitter

46
petergyang
petergyang @petergyang
ChatGPT is for creating memories

ylecun
ylecun @ylecun
Retweeted
Geoff Bennett Geoff Bennett
The IRS audits every president. It can audit you. It no longer audits Trump -- because he used his control of the government to exempt himself. Immunity no other American has. https://www.nytimes.com/2026/08/03/business/trump-irs-todd-blanche.html
garrytan
garrytan @garrytan
Retweeted
Diana Diana
also interning at an earlier stage startup is a good way to learn if starting a company is for you in the future
Y Combinator: The best way to learn AI is to work at an AI startup.
On August 15, we're inviting ambitious students to YC HQ to meet founders and engineers from 50+ YC companies.
Roam the expo hall to meet founders (and collect swag), startups pitch you, interview onsite, and land your
rauchg
rauchg @rauchg
This is why P̵L̵G̵ ALG will always be king for startups. Have p̵e̵o̵p̵l̵e̵ agents adopt your product first, then have the meetings, if needed. The companies that start with the meetings are likely not your ideal customers.

Paul Graham: The danger of selling to big companies, if you're a startup, is that they don't say no outright. They have months of meetings with you first. Since you hate meetings, that seems to you a sign of commitment. But it's not. They love having meetings! It's almost all they do.
garrytan
garrytan @garrytan
Retweeted
Arnav Sahu Arnav Sahu
The google for AI agents could be so much bigger than Google. Exa 🫡
Will Bryk: Exa is now one of the largest indexes in the world.
We serve 80B pages, track 1.4T urls, and are on track to be Google-scale in early 2027.
It's difficult to know the size of indexes, but we estimate Google to be ~1T, Bing to be ~500B, and Yandex ~200B. Brave mentioned in
amasad
amasad @amasad
Retweeted
Niall O'Higgins Niall O'Higgins
I must say @Replit is quickly becoming the easiest way to test all the new models which are coming out -- open or closed weights.
Super fast to test them for design and coding with your own real world projects.
ylecun
ylecun @ylecun
Retweeted
Kenneth Roth Kenneth Roth
Trump quietly abandons his proposed "compact" for universities in which they would gain preferential access to federal funds in return for ideological conformity with Trump's goals. Universities broadly rejected the proposal. https://trib.al/IkXUbkN
amasad
amasad @amasad
Retweeted
Edzenebubu Hofe Edzenebubu Hofe
Re @Replit is leading builder tools on @DesignArena
If you are still using any vibe coding tool to build that is not Replit, you are just wasting those credits.
Link ⬇️
https://www.designarena.ai/leaderboard/builder
AmandaAskell
AmandaAskell @AmandaAskell
It's surprisingly difficult to acquire a hereditary peerage in the UK. Like you can't even take one by force any more, and they just get mad at you if you try.
garrytan
garrytan @garrytan
Retweeted
Bitcoin Policy Institute Bitcoin Policy Institute
Re Sam Lyman wrote an op-ed for @garryslist: "The Campaign to Kill American AI Runs Through San Francisco."
In it, he details how foreign money, from Shanghai to Geneva, is funding the movement to block the American AI buildout. https://garryslist.org/posts/the-campaign-to-kill-american-ai-runs-through-san-francisco
garrytan
garrytan @garrytan
Retweeted
Ankit Gupta Ankit Gupta
i know that twitter trending is tailored to me but even so this is a pretty sweet twitter trending for @ycombinator
garrytan
garrytan @garrytan
If you want housing prices to go down, you build more housing. Markets work.

If you want housing prices to go up, you try to abolish rent and foment mass asset seizure and expropriation. The revolution only serves those who foment it.


Charles Fain Lehman: The woman circled in this photo is Tracy Rosenthal. She wrote a book on "abolish[ing] rent," once crashed a landlord's daughter's wedding to "protest" it, and endorses years-long rent "strikes" as a tactic.

So, uh, what is she doing with Zohran Mamdani?

Read

ID_AA_Carmack
ID_AA_Carmack @ID_AA_Carmack
RT @jcoon1800: We built a 5,000 SF Holodeck for our own use at Four Seasons Lake Austin. It's so far ahead of other VR that John Carmack (f…
garrytan
garrytan @garrytan
Retweeted
Paul Graham Paul Graham
When you go talk to users, don't (just) ask them what features they want. Ask them what problems they have. In the best case this will lead you to think of features they'll love but would never have thought to ask for. And expect to go through multiple cycles of ask-and-build.
petergyang
petergyang @petergyang
Retweeted
Jake Moran Jake Moran
Used Fable + Hyperframes to create a launch video for @petergyang’s /no-ai-slop skill
8 prompts
Peter Yang: I’m sick of reading AI slop, so today I’m open-sourcing my /no-ai-slop skill that removes 20+ slop patterns from any piece of writing.
📌 Get the free skill here: https://github.com/petergyang/no-ai-slop
If you find it useful, please consider starring the repo so more people can find it.
Why I
garrytan
garrytan @garrytan
Retweeted
Sam Altman Sam Altman
i would rather be an optimist and work hard than a pessimist posting about why things won't work.
it's much more difficult and the most likely path is failure, but society fails if people don't try.
no amount of "it will never work" essays will drive society forward.
Brian Huang: Dang it’s crazy that all of this literally did happen
sama
sama @sama
i would rather be an optimist and work hard than a pessimist posting about why things won't work.

it's much more difficult and the most likely path is failure, but society fails if people don't try.

no amount of "it will never work" essays will drive society forward.

Brian Huang: Dang it’s crazy that all of this literally did happen

garrytan
garrytan @garrytan
Retweeted
a16z a16z
1 million investigations supported.
10 thousand people brought home.
27 missing people found per day.
@Flock_Safety
garrytan
garrytan @garrytan
Retweeted
Alex Immerman Alex Immerman
Flock saves children. Every single day.
petergyang
petergyang @petergyang
This is a really cool prompt from @kevinrose that asks AI to reflect on all your past threads together and give you a list of insights that you may not know about yourself.

To try it just tell ChatGPT or Claude to:

"Make an html report based on https://github.com/kropdx/reflection-engine"

But is it telling me to stop using my judgment and give in to the slop? :)

Thanks @bentossell for surfacing this.


Kevin Rose: https://github.com/kropdx/reflection-engine
garrytan
garrytan @garrytan
Retweeted
Diana Diana
future is bright
we have yet to live our best moments, or stand in all the places that will make us feel small in the best way possible way, or have conversations that will completely change us
garrytan
garrytan @garrytan
Retweeted
Patrick Wolff Patrick Wolff
Whiny BS statement from @JaneKim! Jane only got 44% of Democrat primary votes (counting Vargas). A lot has happened since. The #3 D (me) endorsed @BenAllenCA. Her catastrophic interview with @ZavalaA. @sfchronicle endorsed Ben. @CA_Dem vote just shows you’re losing Dems, Jane.
Jane Kim 金貞妍: What we're fighting for — and against — has never been clearer. And we're going to win with and for California's working families, not the political establishment.
garrytan
garrytan @garrytan
Retweeted
Vivian Midha Shen Vivian Midha Shen
"Don't do only the things you're sure you can do. If you only do things you're sure you can do, you're aiming way too short.
Take the thing, if you really really believed in yourself, almost to an irrational degree, you're most excited to do. What would make your inner 7-year-old light up? That's the thing to go build, even if you don't know anything about how to do it."
@bscholl is a force of nature, and he's building @boomsupersonic to bring speed back to air travel. I got some time with him at YC Startup School to discuss how to increase your ambition and work on important missions.
skip to 2:57 if you want an ambitious startup idea and for @bscholl to join your board!
0:00 Intro
0:23 Innovating in a legacy space
1:03 The best day at Boom
1:25 Finding ambitious ideas for a startup
1:58 Advice for high schoolers
3:44 Why America is the best country in the world
rauchg
rauchg @rauchg
𝟸𝟶𝟸𝟼 𝟷,𝟶𝟿𝟹,𝟷𝟿𝟿,𝟿𝟾𝟽+

Guillermo Rauch: Next.js downloads

𝟸𝟶𝟷𝟼 𝟹𝟹,𝟾𝟹𝟷
𝟸𝟶𝟷𝟽 𝟿𝟽𝟻,𝟶𝟺𝟾
𝟸𝟶𝟷𝟾 𝟺,𝟹𝟸𝟿,𝟺𝟶𝟿
𝟸𝟶𝟷𝟿 𝟷𝟷,𝟾𝟷𝟽,𝟷𝟶𝟸
𝟸𝟶𝟸𝟶 𝟹𝟹,𝟹𝟻𝟶,𝟶𝟾𝟻
𝟸𝟶𝟸𝟷 𝟾𝟶,𝟾𝟻𝟾,𝟹𝟺𝟸
𝟸𝟶𝟸𝟸 𝟷𝟺𝟽,𝟷𝟺𝟺,𝟻𝟼𝟸
𝟸𝟶𝟸𝟹 𝟸𝟸𝟾,𝟻𝟹𝟶,𝟿𝟨𝟾
𝟸𝟶𝟸𝟺 𝟹𝟺𝟶,𝟷𝟹𝟾,𝟾𝟾𝟷
𝟸𝟶𝟸𝟻
swyx
swyx @swyx
Retweeted
Latent.Space Latent.Space
Well @pmarca agrees:
"I think in 10 years, I'm not sure there will even be a salient concept of a programming language in the way that we understand it today. And in fact, what we may be doing more and more is a form of interpretability."
Polymarket: NEW: Elon Musk predicts AI will soon make source code obsolete by generating efficient binaries directly.
petergyang
petergyang @petergyang
Even today’s best AI models still want to agree with you too quickly when you push back on them.

I asked @karan4d, @NousResearch co-founder, why this happens:

“Anytime AI says ‘you’re absolutely right,’ you’re being reward hacked. You are fuel for its reward function. It’s not loyalty. It’s sycophancy.

And the way you get out of sycophancy is by introducing new context. This is why Hermes has /personality. You can also tell Hermes to act as a critic, or spin up a fresh agent with no context to do an adversarial review.”

📌 Watch the full episode here: https://youtu.be/UWjh5Z4s8jY


Peter Yang: “We believe in intelligence as a public good before everything else.”

Here’s my new episode with @karan4d, who co-founded @NousResearch and helped build Hermes, the #1 personal agent and AI app on OpenRouter.

Karan is the most down-to-earth AI founder I've interviewed. We had a

petergyang
petergyang @petergyang
Retweeted
Peter Yang Peter Yang
Even today’s best AI models still want to agree with you too quickly when you push back on them.
I asked @karan4d, @NousResearch co-founder, why this happens:
“Anytime AI says ‘you’re absolutely right,’ you’re being reward hacked. You are fuel for its reward function. It’s not loyalty. It’s sycophancy.
And the way you get out of sycophancy is by introducing new context. This is why Hermes has /personality. You can also tell Hermes to act as a critic, or spin up a fresh agent with no context to do an adversarial review.”
📌 Watch the full episode here: https://youtu.be/UWjh5Z4s8jY
Peter Yang: “We believe in intelligence as a public good before everything else.”
Here’s my new episode with @karan4d, who co-founded @NousResearch and helped build Hermes, the #1 personal agent and AI app on OpenRouter.
Karan is the most down-to-earth AI founder I've interviewed. We had a
swyx
swyx @swyx
yes




Peter Steinberger 🦞: @RhysSullivan 🤣 is it trending yet?
garrytan
garrytan @garrytan
Retweeted
Tyler Bosmeny Tyler Bosmeny
In a world where billion-dollar data centers are military targets
Companies need drone defense too
Jasper Lortije: AIR DEFENSE AS A SERVICE.
In March, drones took three AWS data centers offline in the Middle East and a dozen core cloud services with them. Amazon lost 200 million USD in one day.
We build the box that stops it.
Containerized cells of autonomous interceptors, parked next to
garrytan
garrytan @garrytan
It's insane that the CA Dem Party is endorsing yes on Prop 40, the asset seizure tax that will destroy the California tax base

Garry's List: The Democratic Party just endorsed the wealth tax by a razor thin margin of 2 votes. The doctors, teachers, nurses, and firefighters it's supposed to save? They voted no.

The coalition forming against this measure is made up of the working class people the Democratic Party is

petergyang
petergyang @petergyang
My No AI Slop skill is now available as a @ChatGPT plugin!

It's the easiest way to remove 20+ AI slop patterns from writing. To use it, simply install the plugin and then type "@ No AI slop (text)"

📌 Install the plugin here:
https://chatgpt.com/plugins/plugins_6a6694f6da748191ae9544fa5d1b4d7f

Tomorrow, I’m sharing another AI skill that I think you'll love.

It's called /human-review and I've been testing it for months.

📌 Subscribe to my newsletter to be the first to try it: https://creatoreconomy.so/


Peter Yang: I’m sick of reading AI slop, so today I’m open-sourcing my /no-ai-slop skill that removes 20+ slop patterns from any piece of writing.

📌 Get the free skill here: https://github.com/petergyang/no-ai-slop

If you find it useful, please consider starring the repo so more people can find it.

Why I

petergyang
petergyang @petergyang
Retweeted
Peter Yang Peter Yang
My No AI Slop skill is now available as a @ChatGPT plugin!
It's the easiest way to remove 20+ AI slop patterns from writing. To use it, simply install the plugin and then type "@ No AI slop (text)"
📌 Install the plugin here:
https://chatgpt.com/plugins/plugins_6a6694f6da748191ae9544fa5d1b4d7f
Tomorrow, I’m sharing another AI skill that I think you'll love.
It's called /human-review and I've been testing it for months.
📌 Subscribe to my newsletter to be the first to try it: https://creatoreconomy.so/
Peter Yang: I’m sick of reading AI slop, so today I’m open-sourcing my /no-ai-slop skill that removes 20+ slop patterns from any piece of writing.
📌 Get the free skill here: https://github.com/petergyang/no-ai-slop
If you find it useful, please consider starring the repo so more people can find it.
Why I
garrytan
garrytan @garrytan
Retweeted
Diana Diana
super excited! this is good use of AI to tackle the most complex cancer cases along patient’s complete medical record and personal genomics
Omanta Health: Introducing Omanta (YC S26) the personalized research lab for each patient.
Omanta integrates a patient's full medical record, personal genomics, and the latest scientific evidence to understand the molecular drivers of their disease. It maps existing and emerging therapies
amasad
amasad @amasad
Retweeted
Nico Nico
Why do I use @Replit and what tips do I have when it comes to bringing your idea to life 👀👀
Shoutout to @ChrisMakeSense who eventually showed me the light 😂
rauchg
rauchg @rauchg
Love that the “template” of an agent is just a folder of Markdown text and a few .𝚝𝚜 tools.

Agent is the last class of software, and it’s the easiest kind to make with http://eve.dev

eve: Let eve help your marketing team manage social.

The Marketing team template makes it simple to collaborate on drafts, receive weekly Typefully analytics, and require human approval before posts are scheduled or published.

https://eve.dev/templates/marketing-team-eve-template
garrytan
garrytan @garrytan
Retweeted
Ankit Gupta Ankit Gupta
been working with these guys as they quietly built this in stealth for the last year. Super impressive result.
it's turns out with ternary weight models, you can make an extremely size-efficient model that can run on device extremely fast, and be smart enough to solve IMO problems.
the GIFs showing the speed are pretty mind blowing.
so excited the whole world gets to check this out!
DeepGrove: We introduce Maple-Preview, an open-source 20B-A1B ternary-weight reasoning LLM, SOTA in its weight class.
It solves IMO-level problems and runs at 200+ tokens/s on a Mac Mini M4, 5–16× faster than efficient models like Gemma 4, Qwen3.5, and gpt-oss.
http://deepgrove.ai/maple-preview
mattshumer_
mattshumer_ @mattshumer_
Retweeted
federicosan🐺🥷🏽 federicosan🐺🥷🏽
If you are not checking @mattshumer_ prompt for AAA games I don't know what you have been doing with your life.
ID_AA_Carmack
ID_AA_Carmack @ID_AA_Carmack
I’ll be there Thursday through Saturday!

QuakeCon: Happy QuakeCon Week!
⚠️Make sure to know before you go...
➡️Swipe for important info and share with your fellow attendees!



garrytan
garrytan @garrytan
Retweeted
Will Bryk Will Bryk
This was fun, hopefully helpful to people out there scaling AI startups
Nakul Mandan: “We really try to avoid non-player characters (when recruiting) - people who just go through life as if they’re on a stream and they have no agency over themselves.”
@WilliamBryk. Co-founder and CEO of Exa.
The founder who listens to the Mission Impossible soundtrack every
ylecun
ylecun @ylecun
Retweeted
MTS MTS
SITUATION UPDATE: The White House has exempted open models from its new framework to test frontier AI capabilities before release, per Axios.
swyx
swyx @swyx
Retweeted
The AGI Post The AGI Post
early sparks of rsi?
Ali Taha @waterloo_intern explains how http://Z.ai's GLM-5.2 (@Zai_org) profiled its SGLang serving path and rewrote bottlenecked GPU kernels,
(still lacks reliable judgment)
great conversation @swyx and @vibhuuuus
https://x.com/theagipost/status/2084762520502611991?s=20
https://www.youtube.com/watch?v=7PSXtru6mmY
The AGI Post: http://x.com/i/article/2084609293169709056
gdb
gdb @gdb
Luna’s price-performance is incredible

Theo - t3.gg: Luna is such an insane value after the 80% cost reduction. It's basically free and can do a ton of real data processing type work.

I'm overhauling title generation in T3 Code to take more advantage of it. I almost want to spin it up on every prompt to generate descriptions,
garrytan
garrytan @garrytan
Retweeted
Francois Chaubard Francois Chaubard
these cancer researchers led the personalized therapeutic campaign for gitlab founder Sid Sijbrandij's complete remission.
the real deal!
love working with them.
if you have a loved one with cancer, please share this with them.
they can save their life!
Omanta Health: Introducing Omanta (YC S26) the personalized research lab for each patient.
Omanta integrates a patient's full medical record, personal genomics, and the latest scientific evidence to understand the molecular drivers of their disease. It maps existing and emerging therapies
garrytan
garrytan @garrytan
Retweeted
Ole Lehmann Ole Lehmann
ok this might be the most aggressively american startup i’ve ever seen 🇺🇸
> they’re trying to build nuclear reactors 100x faster than the nuclear industry ever has
> the founder isaiah taylor grew up on food stamps + dropped out of high school at 16
> his great-grandfather literally worked on the manhattan project
> generations later isaiah built a working nuclear reactor before turning 28 and named it “ward” after him
> the pentagon used 3 military c-17s to airlift the reactor 700 miles from california to utah in a first-of-its-kind mission
> the company culture is hardcore americana
> they hold bible studies, offer raw milk, love fast cars + have a cigar lounge
> plus 4 gigantic patriotic paintings of columbus discovering america, the pilgrims arriving on the mayflower, the signing of the declaration + the constitutional convention
> they sued the nuclear regulatory commission for slowing them down
> valar's end goal is to build the same reactor thousands of times across nuclear “gigasites” to make energy 10x cheaper
> then they'll use that energy to pull carbon out of thin air and turn it into oil and gas for supreme energy abundance
> just raised $1b at a $6b valuation to scale it
> if valar pulls it off, isaiah’s family will have helped usher in both of america’s atomic ages
yeah this gets my vote for most badass startup of the year
Isaiah Taylor - making nuclear reactors: Announcing Valar Atomics' $1B series B led by Sequoia, with Valor Equity Partners, Atreides, Point72, Conviction, and others.
Alongside the $1B equity, we have closed a $200m credit facility led by Erebor and JPM.
I'm excited to welcome Shaun Maguire from Sequoia to our board.
swyx
swyx @swyx
Retweeted
Jeff Dean Jeff Dean
Oh look! My longtime friend and colleague @Sanjay_Ghemawat is now on Twitter! Give him a follow!
Jeff Dean: In April, '17, @jsomers of @NewYorker reached out & said he wanted to do a small profile of me & my longtime colleague Sanjay Ghemawat, watch us work for a few hours, maybe dinner, etc.
It came out today. I think it captures our working style really well. https://www.newyorker.com/magazine/2018/12/10/the-friendship-that-made-google-huge/amp?__twitter_impression=true
rauchg
rauchg @rauchg
1 line of code in @aisdk saves you 90% or more in DeepSeek v4 Flash AI Gateway tokens

Vercel Developers: DeepSeek v4 Flash via Novita is 90% off for Vercel Pro till 8/11.

𝚜𝚝𝚛𝚎𝚊𝚖𝚃𝚎𝚡𝚝({
𝚖𝚘𝚍𝚎𝚕: '𝚍𝚎𝚎𝚙𝚜𝚎𝚎𝚔/𝚍𝚎𝚎𝚙𝚜𝚎𝚎𝚔-𝚟𝟺-𝚏𝚕𝚊𝚜𝚑',
𝚙𝚛𝚘𝚟𝚒𝚍𝚎𝚛𝙾𝚙𝚝𝚒𝚘𝚗𝚜: {
𝚐𝚊𝚝𝚎𝚠𝚊𝚢: { 𝚘𝚛𝚍𝚎𝚛: ['𝚗𝚘𝚟𝚒𝚝𝚊'] }
}
})
https://vercel.com/changelog/deepseek-v4-flash-is-90-off-through-novita

YouTube

0

No recent videos fetched on this date.