bat

788 posts

bat banner
bat

bat

@baptiste_cumin

https://t.co/O5kgtLv2TE

Katılım Ocak 2017
971 Takip Edilen239 Takipçiler
Luke Metro
Luke Metro@luke_metro·
people laugh at New Media but Cognition has successfully turned videos of the CEO doing math into $26B in shareholder value
English
2
0
7
530
bat
bat@baptiste_cumin·
@carlrivera @Shopify Congrats Carl! You set a high bar starting my career working with you. Admire what you built at Shop and good luck with the next big thing
English
0
0
0
281
Carl Rivera
Carl Rivera@carlrivera·
I'm moving on from @Shopify. When I joined 7.5 years ago, friends were literally taking bets on whether I'd last a year. The idea of me thriving inside a large company (and Shopify was a lot smaller back then!) felt unlikely. Most people assumed I'd be itching to start my next thing. Instead, Shopify became my company. I wasn't a founder, but @tobi always made me feel like one. It wasn't a startup, but Shopify moves like one. And the mission isn't just words — it shows up in the decisions, every day. I got to wear a lot of hats. Introducing Shop to the world. Going deep on the rails of how money moves on the internet. Reimagining Shopify for the age of AI through design. Leaving is hard. Shopify changed my life in more ways than I can recount. It taught me how to build for the long term, gave me a bar for craft and ambition I'll carry into everything that comes next, and showed me what it looks like when a company lives its mission. In my next chapter, I'm getting the opportunity to bring together all the things I've obsessed over the past decade — consumer mobile, finance and money movement, design, AI — into a single role. More on that soon.
English
118
11
1.2K
145.5K
bat
bat@baptiste_cumin·
@yoavgo @mgalle I don't think conferences work like this in practice, we've sponsored in the last week before an event in the past
English
1
0
0
306
(((ل()(ل() 'yoav))))👾
@mgalle the list of sponsors (and hence the amount of available sponsorship money) should be known and available *before* setting the registration prices...
English
2
0
3
1.2K
Michael J. Miraflor
Michael J. Miraflor@michaelmiraflor·
If you're an aspirational Tech podcast, this now former list of TBPN sponsors probably has dollars to re-allocate ASAP. Hop on the phone and close those deals!
Michael J. Miraflor tweet media
English
45
31
802
106.6K
Joe Weisenthal
Joe Weisenthal@TheStalwart·
This is the schism of the left right here. By having staked the claim early on, that LLMs are a gimmick or a "stochastic parrot" that can't actually know anything, it's unacceptable, from the perspective of many, to actually take seriously the capabilities of the models.
English
36
57
634
116.5K
bat
bat@baptiste_cumin·
@chaykak ai design tool framer
bat tweet media
English
0
0
1
393
Kyle Chayka
Kyle Chayka@chaykak·
this week's column is for all my taste bros out there
Kyle Chayka tweet mediaKyle Chayka tweet media
English
32
58
1K
455.6K
bat
bat@baptiste_cumin·
@DKThomp Wrong, tokens = outputs + inputs. Most costs are inputs: web research (junior consultants) would cost more than devs on this. Devs just drive most revenue because all their work context is readily digitized and available
English
0
0
0
8
Derek Thompson
Derek Thompson@DKThomp·
New newsletter: The transcript of my AI bubble conversation, with @pkedrosky. Feat.: - Why did the Mag7 equity miracle suddenly stop? - The growing private credit crisis, explained - Why the enormous revenue boom from new agents like Claude Code might be a sugar high, in which explosive revenue growth today precedes much slower revenue growth after AI adoption among software engineers peaks - Where equity value is flowing if it’s leaving software - Why US productivity seems to be rising but actually isn't derekthompson.org/p/yes-ai-is-a-…
English
9
8
87
104.2K
bat
bat@baptiste_cumin·
@juliarturc Are diffusion models making a comeback? Seeing lots of diffusion content lately but i'm out of the loop since nano-banana era autoregressive image gen
English
1
0
1
406
Julia Turc
Julia Turc@juliarturc·
When reading diffusion papers, my most common reaction is "mkay the math works out but WHY WOULD A SANE PERSON CHOOSE TO DO THIS". This included the (foundational) DDPM formulas. The saving grace is that you can visualize it as 2D particle motion and get a solid intuition.
English
11
15
185
20.5K
Shakeel
Shakeel@ShakeelHashim·
There is a section in this where @dwarkesh_sp absolutely runs rings around Elon's understanding of alignment and the like. It's very revealing — left with the impression that Elon really doesn't know what he's talking about.
John Collison@collision

Please enjoy this Cheeky Pint / @dwarkesh_sp crossover with @elonmusk. Dwarkesh was most interested in how Elon is going to make space datacenters work. I was most interested in Elon's method for attacking hard technical problems, and why it hasn’t been replicated as much as you might expect. But we got into plenty of topics in this three-hour session. 00:00:23 Space GPUs 00:35:39 Alignment 00:58:48 xAI 01:15:01 Optimus 01:28:03 China 01:40:46 Management 02:16:38 DOGE 02:34:58 Space GPUs redux

English
35
16
286
63.3K
bat
bat@baptiste_cumin·
@Ronanchamberss Is there a single video online of sam altman laughing
English
0
0
0
17
Ronan
Ronan@Ronanchamberss·
I do not think he laughed.
Ronan tweet media
Sam Altman@sama

First, the good part of the Anthropic ads: they are funny, and I laughed. But I wonder why Anthropic would go for something so clearly dishonest. Our most important principle for ads says that we won’t do exactly this; we would obviously never run ads in the way Anthropic depicts them. We are not stupid and we know our users would reject that. I guess it’s on brand for Anthropic doublespeak to use a deceptive ad to critique theoretical deceptive ads that aren’t real, but a Super Bowl ad is not where I would expect it. More importantly, we believe everyone deserves to use AI and are committed to free access, because we believe access creates agency. More Texans use ChatGPT for free than total people use Claude in the US, so we have a differently-shaped problem than they do. (If you want to pay for ChatGPT Plus or Pro, we don't show you ads.) Anthropic serves an expensive product to rich people. We are glad they do that and we are doing that too, but we also feel strongly that we need to bring AI to billions of people who can’t pay for subscriptions. Maybe even more importantly: Anthropic wants to control what people do with AI—they block companies they don't like from using their coding product (including us), they want to write the rules themselves for what people can and can't use AI for, and now they also want to tell other companies what their business models can be. We are committed to broad, democratic decision making in addition to access. We are also committed to building the most resilient ecosystem for advanced AI. We care a great deal about safe, broadly beneficial AGI, and we know the only way to get there is to work with the world to prepare. One authoritarian company won't get us there on their own, to say nothing of the other obvious risks. It is a dark path. As for our Super Bowl ad: it’s about builders, and how anyone can now build anything. We are enjoying watching so many people switch to Codex. There have now been 500,000 app downloads since launch on Monday, and we think builders are really going to love what’s coming in the next few weeks. I believe Codex is going to win. We will continue to work hard to make even more intelligence available for lower and lower prices to our users. This time belongs to the builders, not the people who want to control them.

English
1
0
28
2.2K
bat
bat@baptiste_cumin·
@patrickc @brian_armstrong Assuming giving 3.5% yield to their depositors won't bring them tons of users and allow them to compete with banks, it's principled
English
0
0
1
355
Patrick Collison
Patrick Collison@patrickc·
Something this WSJ piece omits is that what @brian_armstrong is arguing for is not obviously in Coinbase's interest: in a world where yield sharing is prohibited, USDC will at least in a first-order way be more directly profitable for Coinbase. As far as I can tell, Brian's position stems from his belief in the importance of a vibrant and competitive market that protects consumer liberty, and I think he deserves a lot of credit for that. wsj.com/finance/curren….
English
143
226
2.8K
455.3K
bat
bat@baptiste_cumin·
@eric_seufert Merchants can turn it off, and most already pay 30%+ on ads. IMO the reason it won't take off is just that most purchases are not made via chat
English
0
0
0
120
Eric Seufert
Eric Seufert@eric_seufert·
Two days ago, OpenAI clarified the transaction fee it will charge Shopify merchants with its Instant Checkout product: 4%. This reinforces my argument that (independent) agentic commerce is a mirage: many Shopify merchants run on incredibly thin margins (3-8% net) and simply may not be able to support this. Further, they aren't in control of it: if ChatGPT's Instant Checkout affiliate link system overwhelms or front-runs their existing organic discovery, it could be disastrous for their businesses. Compare this to an on-platform shopping agent like Amazon's Rufus or Walmart's Sparky, which benefits in a number of ways when a customer purchases through their platform, obviating the need for the agent to charge a direct transaction fee. This transaction fee stacks on top of the existing merchant and payment processing fees that retailers pay. For CPG retailers with fixed COGS, an additional, obligatory 4% fee on an unpredictable volume of sales could fundamentally compromise the durability of their business. It also highlights the value of advertising relative to Instant Checkout's affiliate model: advertisers control the margin they forfeit through advertising with their bid and can optimally price it. For many retailers, 4% may be unmanageable as a transaction fee, but some lower relative amount could be feasible in the form of a bid. This is the difference between endogenous pricing and an exogenous imposition of friction.
Eric Seufert tweet media
English
32
33
361
117.2K
bat
bat@baptiste_cumin·
@_xjdr The biggest jump was the move from collab filtering for candidates to two tower models in candidate generation. The feature generator there also seems to be the black box and the core of the model.
English
0
0
2
516
xjdr
xjdr@_xjdr·
many many years ago, i worked on recsys and personalization systems. i finally looked at the xai algorithm and i do not understand it at all. its the normal candidate generation, pruning, reranking but without the core parts like collaborative filtering / matrix factorization and any type of simclustering like functionality seems missing? i didnt spend that much time looking at it but the best i could tell grok does the feature engineering in a black box (which is where all the magic is) and they arent using any sort of persona graph anymore? its very possible (even likely) that i am just way behind the SOTA after all these years but from where im sitting this seems like a bad design (even if the implementation is perfect)
English
19
12
309
25.1K
bat
bat@baptiste_cumin·
@GabLattanzio La "recherche" qu'il cite est un rapport opaque publié par une entreprise qui vend des services de fiabilité aux boites IA
Français
0
0
0
171
Gabriel Lattanzio
Gabriel Lattanzio@GabLattanzio·
"Vous savez messieurs, l'appareillage de monsieur Gutenberg est d'une imprécision incroyable, vous n'imaginez pas le nombre de pages gâchées, et l'impossibilité de nuancer la forme d'une majuscule." - Lukas Julianus en 1453.
Fabien@Fabien_Mikol

Alors aujourd'hui Luc Julia a été carrément auditionné à l'Assemblée nationale. Ce fut donc pour lui l'occasion de répéter à nouveau les mêmes absurdités sur le "taux de pertinence de Hong-Kong", la dégradation chaque année, l'étude Newsguard mal comprise, etc. C'est une folie.

Français
7
47
370
17.6K
bat
bat@baptiste_cumin·
@Lau_hed @Fabien_Mikol Non 1. Les questions ne sont pas open-source (10 questions tirées au sort de leur dataset privée) 2. C'est une démarche commerciale, non académique: Newsguard vend un produit dans ce domaine 3. Toutes les autres recherches publiées vont dans l'autre sens
Français
1
0
1
17
Fabien
Fabien@Fabien_Mikol·
Alors aujourd'hui Luc Julia a été carrément auditionné à l'Assemblée nationale. Ce fut donc pour lui l'occasion de répéter à nouveau les mêmes absurdités sur le "taux de pertinence de Hong-Kong", la dégradation chaque année, l'étude Newsguard mal comprise, etc. C'est une folie.
Français
49
57
428
266.5K
bat
bat@baptiste_cumin·
@jujulemons It took 9 months for us to get a founder visa. The reason? Home Office was suspicious of us having VCs backing us from the US. 🤦‍♂️
English
0
0
0
24
Julia Willemyns
Julia Willemyns@jujulemons·
Recommendation 4: pilot new Global Talent endorsers Global Talent is meant to be the UK’s “best and brightest” route. In practice, it’s skewed towards academia. Home Office evaluation: 67% of Global Talent respondents were in academia. That makes sense given how endorsement works today. It’s easier to evidence “exceptional talent” with publications, citations, grants, prizes, and institutional affiliations. And most endorsers are academic in nature. But a lot of genuinely exceptional applied talent doesn’t look like that on paper: 1. founders building frontier products under NDAs 2. senior engineers whose best work is proprietary 3. people with major industry impact but fewer traditional “signals” (papers, prizes) I propose a 12–24 month pilot: add ~5 additional endorsers outside academia (e.g. leading VC funds / major R&D organisations), give each a clear visa cap, publish criteria, and evaluate outcomes (earnings, innovation activity, retention). Keep what works; scrap what doesn’t.
English
3
2
27
1.7K
Julia Willemyns
Julia Willemyns@jujulemons·
How can we MAKE BRITAIN WIN the great global talent war. The UK already has two strong routes for exceptional talent. They have strong outcomes but are underused (~3% of work visas combined). I have come up with four quick fixes that would increase uptake of exceptional talent routes by the SMARTEST MOST TALENTED people. Big caveats up front: 1. I don’t think we should lower standards or “water down” who gets in. We shouldn’t. The point is to keep the bar high, and fix the parts of the system that stop genuinely exceptional people from choosing Britain. 2. People move for lots of reasons: opportunity, taxes, safety, schools, weather. The UK is competitive on some of these and not on others - but there’s one thing we can improve quickly: the visa offer. 3. It’s not true that the public would be automatically hostile to this. When you ask about highly skilled / top talent migration, people are much more supportive than the general immigration debate suggests, especially when it’s framed around contribution.
Julia Willemyns tweet media
English
11
25
153
34.8K
bat
bat@baptiste_cumin·
@birk @shop Teaching Canadians about Crayfish parties has to be in the achievement list somewhere
English
1
0
1
155
Birk Jernström
Birk Jernström@birk·
It's a forgotten story, but yes. I co-founded Tictail in 2011; an e-commerce platform and marketplace. Acquired by Shopify in 2018 where we built @shop Our biggest innovation at Tictail? Money compression. Raised $32M. Sold for $17M. AMA.
Jonathan Grahl@jonathangrahl

@birk You made Tictail? Thats dope

English
14
4
305
63K