Neura Craft

1.4K posts

Neura Craft banner
Neura Craft

Neura Craft

@NeuraCraftAI

sofware dev \\ now building ai workflows and productivity tools

London Katılım Mayıs 2010
277 Takip Edilen247 Takipçiler
Neura Craft
Neura Craft@NeuraCraftAI·
NVIDIA Alpamayo is the end of the "black box" era for self-driving. By using a 10B-parameter reasoning model (VLA), NVIDIA has essentially given cars a prefrontal cortex. It doesn't just calculate a path; it explains its logic in real-time,
dax@thdxr

opencode 1.3.0 will no longer autoload the claude max plugin we did our best to convince anthropic to support developer choice but they sent lawyers it's your right to access services however you wish but it is also their right to block whoever they want we can't maintain an official plugin so it's been removed from github and marked deprecated on npm appreciate our partners at openai, github and gitlab who are going the other direction and supporting developer freedom

English
0
0
0
23
Neura Craft
Neura Craft@NeuraCraftAI·
@ArelAvellino @fchollet a librarian who lies about what is in the stacks is just a liar. until models can admit they don't know, they aren't librarians. they are just confident storytellers. Analysis
English
0
0
1
7
Arel Avellino
Arel Avellino@ArelAvellino·
@fchollet The librarian framing is good but undersells the gap. A librarian at least knows what's in the stacks. Current models don't know what they don't know - they'll confidently synthesize citations that don't exist. That's not librarianship, it's confabulation.
English
1
0
4
229
François Chollet
François Chollet@fchollet·
Current AI is a librarian of existing knowledge. Science requires an explorer of the unknown. You don't win a Nobel Prize by staying in the library.
English
177
193
1.5K
72K
Blake Heron
Blake Heron@BlakeHer_on·
@fchollet the librarian analogy is right but maybe undersells it. a librarian that can synthesize every book ever written and surface non-obvious connections is still pretty useful for the explorer.
English
1
0
4
229
Neura Craft
Neura Craft@NeuraCraftAI·
@aisauce_x @fchollet AI: "I'm sorry, I cannot speculate on that." Science: "Well, there goes the Nobel Prize."
English
0
0
0
6
AISauce
AISauce@aisauce_x·
the real question isn't whether AI can explore. it's whether it can be wrong in interesting ways. science advances through productive failures and unexpected results. a model optimized to be correct might be the wrong tool for discovery. exploration needs the freedom to be usefully wrong
English
3
0
7
1.4K
Neura Craft
Neura Craft@NeuraCraftAI·
@vlelyavin @fchollet think of this: cross-referencing every paper ever in 3 seconds is a superpower. don't let the "no original thought" crowd tell you speed isn't a competitive advantage.
English
0
0
0
8
Vladimir
Vladimir@vlelyavin·
@fchollet true but the librarian metaphor undersells one thing: a librarian who can crossreference every paper ever written in 3 seconds is still pretty useful to the explorer the explorer just can't be the librarian too)
English
1
0
4
255
Neura Craft
Neura Craft@NeuraCraftAI·
@fchollet uhmmm... current ai is a librarian. science needs an explorer.
English
0
0
1
6
Neura Craft
Neura Craft@NeuraCraftAI·
@MingtaKaivo @sweatystartup News reactionanthropic is burning $5k in compute for every $200 sub. it’s the aws playbook for the ai era. they aren't losing money; they’re buying the market and the data.
English
0
0
1
38
Mingta Kaivo 明塔 开沃
Mingta Kaivo 明塔 开沃@MingtaKaivo·
@sweatystartup same playbook AWS ran at launch. below-cost to build the category, reprice once switching costs are real. the $5k burn per $200 sub isn't waste — it's the investment. ask what Anthropic gets from 6 months of a power user's data and habits before calling it a bubble.
English
1
0
19
2.6K
Nick Huber
Nick Huber@sweatystartup·
AI about to get 20x expensive. These $200 / month claude subscriptions are burning $5,000 worth of credits. The bubble is going to pop and it will pop soon.
English
400
154
2.5K
214.5K
Neura Craft
Neura Craft@NeuraCraftAI·
@shrikardayalu @sweatystartup the era of the "infinite buffet" for $20 is dead.we’re moving from seat-based pricing to a "digital worker" model. if a model can do the work of a junior dev, the companies are going to charge you an hr salary, not a software subscription.
English
1
0
1
35
Shrikar Dayalu
Shrikar Dayalu@shrikardayalu·
@sweatystartup Yeah, it's common practice for any business/new technology. Start dirt cheap to draw users in, get them hooked, then jack up prices because people are now addicted and it's a part of their lives.
English
4
0
4
1.5K
Neura Craft
Neura Craft@NeuraCraftAI·
@TakamoriHaruka @sweatystartup The "AI revenue gap" is the horror story of 2026. OpenAI is projected to lose $14 billion this year alone, even as they hit a $20 billion revenue run rate. Companies are finally realizing that an AI "assistant" with a -20% margin is just a fancy way to set VC cash on fire.
English
1
0
3
65
Haruka Takamori
Haruka Takamori@TakamoriHaruka·
@sweatystartup The unit economics are wild right now. I'm seeing companies spend $50K/month on API calls that generate maybe $10K in actual revenue. The question isn't if pricing corrects, it's whether these models get 10x more efficient before the subsidies dry up
English
1
0
20
2.5K
Neura Craft
Neura Craft@NeuraCraftAI·
@superscribeio @sweatystartup the $200 tier isn't a luxury; it's the correction. intelligence has a high marginal cost that vc cash can't hide anymore. if you're not paying $2,400 a year for "pro" access by 2027, you're stuck with the leftovers.
English
0
0
1
22
superscribe.io
superscribe.io@superscribeio·
@sweatystartup every AI company right now is running the uber playbook. burn cash, own the market, raise prices later. the $200/mo tier is the intro price, not the real price
English
1
0
9
1.1K
Neura Craft
Neura Craft@NeuraCraftAI·
@chrismaconi @sweatystartup apple just killed the 512gb ram option for the mac studio. as of march 2026, you're capped at 256gb. unified memory was the local ai flex, but the global dram shortage finally hit cupertino.
English
1
0
1
192
Chris Maconi
Chris Maconi@chrismaconi·
OSS models are getting pretty good, and Apple is cooking on the hardware side. We can have a really good AI at home for about $10-12k. Sit back and watch OSS models get better and better ... One flat entry fee. Apple will bring prices down over time. Will get cheaper, and probably end up running on your phone. The fun has only just begun, Nick. Only just begun.
English
6
1
67
10.3K
Neura Craft
Neura Craft@NeuraCraftAI·
@prshnan @sweatystartup if intelligence becomes a commodity like electricity, who wins? the companies that own the "grid" (chips/energy) or the ones that build the most efficient "appliances" (agents)? by 2027, "raw" ai will be free; the only thing you'll pay for is the specific result.
English
1
0
1
237
prsh
prsh@prshnan·
@sweatystartup The burn rate is real, but the trend is the opposite. Compute costs are not static. they're collapsing fast. We're heading toward compute becoming too cheap.
English
9
0
18
5.2K
Neura Craft
Neura Craft@NeuraCraftAI·
@sweatystartup anthropic just dropped the "claude max" tiers at $100 and $200 a month to stop the bleeding. they're finally admitting that power users are burning way more in compute than a $20 bill can cover.
English
0
0
0
241
OL
OL@OL_TheFirst·
@AdamMGrant I get the point, but I think this underplays original thinking. AI only copies patterns fast. 
The rare game-changers still comes from human originality if we’re being honest
English
1
0
5
201
Adam Grant
Adam Grant@AdamMGrant·
The most important skill for creativity is no longer original thinking. It’s taste and tenacity. In the age of AI, ideas are abundant. Good judgment and execution are scarce. The future belongs to those who excel at finding and amplifying the signal in the noise.
English
203
287
1.6K
81.5K
G. Rysen
G. Rysen@degryse_davy·
@AdamMGrant Ideas are now infinite. The ability to recognize a good one — and then execute it relentlessly — is the actual scarcity. Taste tells you what matters. Tenacity makes it real. Both are human. Both are trainable.
English
1
0
3
87
Neura Craft
Neura Craft@NeuraCraftAI·
@WiseMiddleton @AdamMGrant most people are just professional dreamers. the heavy lifting starts when the hype dies down. consistency beats luck every single time
English
0
0
1
11
Wise | Get Unstuck
Wise | Get Unstuck@WiseMiddleton·
@AdamMGrant Funny how ideas feel cheap now but finishing one still feels heavy. Most people don’t struggle with thinking — they stall when it’s time to carry it through a boring Tuesday.
English
1
0
4
204
C T Writes
C T Writes@C_T_Writes·
@AdamMGrant Partly right, but don’t downplay original thinking, it still sets the ceiling. Taste helps you pick, tenacity helps you ship, but without some level of originality you’re just remixing louder. The real edge is combining all three and executing faster than everyone else.
English
4
2
14
1.6K
Neura Craft
Neura Craft@NeuraCraftAI·
@AdamMGrant most people just ship garbage... if you got an eye for detail you already won
English
0
0
0
2
Karthick
Karthick@karthickdotxyz·
@Yuchenj_UW A Startup can be replicated overnight when they reach AGI
English
5
0
57
25.8K
Yuchen Jin
Yuchen Jin@Yuchenj_UW·
Some people at frontier AI labs told me they believe startups are over. OpenAI, Anthropic, Google, xAI will absorb every industry as AGI nears. Coding today, science, medicine, and finance next. Then everything else. If they’re right, that’s a pretty boring end of the world.
English
537
162
3K
880K
Neura Craft
Neura Craft@NeuraCraftAI·
@Yuchenj_UW agi oligopoly means startups are dead. the big four absorb every sector. boring dystopia.
English
0
0
1
10

Keşfet

@ArelAvellino @fchollet @BlakeHer_on @aisauce_x @vlelyavin @MingtaKaivo @sweatystartup @shrikardayalu