N3crafter

561 posts

N3crafter

N3crafter

@n3crafter

too much I want to learn not enough time learn anyways

Katılım Temmuz 2024
184 Takip Edilen17 Takipçiler
N3crafter
N3crafter@n3crafter·
@thdxr A production product, sure. Internal or personal tools/products not so much
English
0
0
0
174
dax
dax@thdxr·
think back to projects you've worked on in the past it's hard not to imagine they'd have been completed way faster now that we have ai but everything still feels as slow and as difficult as ever
English
106
33
1.2K
62.8K
N3crafter
N3crafter@n3crafter·
@zeeg Agree but only post Opus 4.5
English
0
0
0
55
David Cramer
David Cramer@zeeg·
Whenever someone talks about how much models have improved over the last 3, 6, or 12 months I’m like “sure, but not enough to matter”. I’m still solving the same problems I was solving a year ago and there’s no plausible path forward. If you honestly ask yourself how much has genuinely changed I think it will make you a lot more grounded about how much might change in the future. It’s not to say there’s not been visible improvement, but there have been no exponential leaps in capabilities of the technology. Only exponential micro benchmarks.
Nate Berkopec@nateberkopec

x.com/i/article/2058…

English
34
12
325
98.2K
N3crafter
N3crafter@n3crafter·
@thsottiaux When you say efficient, can we actually expect limits to increase over time but it seems the trend is for limits to keep going down
English
0
0
1
139
Tibo
Tibo@thsottiaux·
Our master plan is to release better and more efficient models. And also to release better products, week after week. Oh and get more compute too. Together with spending too much time on x. How good is this plan?
English
533
86
4.3K
217.5K
N3crafter
N3crafter@n3crafter·
@brotzky I love anything performance. keep it coming
English
0
0
1
53
Brotzky
Brotzky@brotzky·
Already talking to a bunch of people about performance issues and how they solved them! Turns out people are kind, awesome, and want to help. This is all for fun (no promises on timelines) but I’m excited to write more asap.
Brotzky@brotzky

Introducing performance.dev! A new space where I explore how the best apps in the world are built. First piece: How's Linear is so fast? a technical breakdown. performance.dev/how-is-linear-…

English
11
4
164
14.1K
N3crafter
N3crafter@n3crafter·
@fjzeit Yes, if you want people to try and use lode coding, videos would help a lot
English
0
0
1
20
fj
fj@fjzeit·
@n3crafter no but i have been asked before. i should make some.
English
2
0
2
53
fj
fj@fjzeit·
you'll all be lode coding or something similar within 6 months...
Hedgie@HedgieMarkets

🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an internal memo warning the company burned through its entire 2026 AI budget in just four months. American AI software prices have jumped 20% to 37%, and GitHub (owned by Microsoft) is dropping flat-rate plans for usage-based billing across its products. My Take The AI subsidy era is ending in real time. The same company that put $13 billion into OpenAI and built the Azure infrastructure powering most of Anthropic's compute just looked at the bill from a competitor's coding tool and decided it was not worth paying. That is not a productivity failure on Anthropic's end. Token-based pricing is forcing every enterprise customer to confront the actual cost of running these models at scale, and the number turns out to be far higher than the flat-rate experiments suggested. This ties directly to my Gemini Flash post yesterday. Anthropic, OpenAI, and Google all raised effective prices in the last six months. Enterprises that built workflows assuming AI costs would keep falling are now watching annual budgets evaporate in months. Two outcomes look likely from here. Either enterprises scale back AI usage to fit budgets, which slows the revenue ramp the labs need to justify their valuations ahead of IPOs, or the labs cut prices and absorb the losses, which makes the unit economics worse at exactly the wrong moment. Both paths land in the same place, the numbers stop working, and somebody has to take the writedown. Hedgie🤗

English
5
1
17
2.1K
Sam Altman
Sam Altman@sama·
what problem do you most hope AI will solve in the future? maybe we can help!
English
14.9K
734
12.6K
3.5M
N3crafter
N3crafter@n3crafter·
@burkeholland That’s what I’ve been saying. No major leap since Opus 4.5, at least not where it matters.
English
0
0
1
140
Burke Holland
Burke Holland@burkeholland·
It seems to me that significant improvements in models won’t happen until we discover a method other than simply increasing the number of parameters. This is why it feels like we kinda hit a ceiling with Opus 4.5. Cause we did.
English
21
0
48
5.2K
Demis Hassabis
Demis Hassabis@demishassabis·
I’ve always believed the No.1 application of AI should be to improve human health. That work started with AlphaFold, and now at @IsomorphicLabs with the mission to reimagine drug discovery and one day solve all disease! We are turbocharging that goal with $2.1B in new funding.
English
704
2.7K
21.3K
3.1M
Sam Altman
Sam Altman@sama·
@icanvardar curious to see if you still feel this way after the next model!
English
212
39
2.6K
214.9K
Can Vardar
Can Vardar@icanvardar·
gpt 5.5 is already enough for most programming work the bottleneck is no longer the model
English
102
78
2.4K
229.1K
Sam Altman
Sam Altman@sama·
what would you most like to see improve in our next model?
English
8.3K
305
9K
1.4M
Mario Zechner
Mario Zechner@badlogicgames·
but it's cool that frontier models are now basically regressing. maybe all this madness will come to an end soon.
English
28
13
391
34K
N3crafter
N3crafter@n3crafter·
@thdxr how long until you are compute constrained...?
English
0
0
0
207
N3crafter
N3crafter@n3crafter·
@thdxr Is this one accurate -> we are subsidizing the research from AI labs?
English
0
0
0
211
dax
dax@thdxr·
the weirdest thing that's happened working in inference is understanding how much people are making things up i see posts every single day about inference, costs, subsidization etc it's all just theories that sound right but are completely wrong
English
41
12
709
42.8K
Mario Zechner
Mario Zechner@badlogicgames·
jfc imma stop using agents for the rest of the week. garbage.
Mario Zechner tweet media
English
12
1
87
13.1K
N3crafter
N3crafter@n3crafter·
@sama We need 2x for Plus until next year 🙏
English
0
0
0
13
Sam Altman
Sam Altman@sama·
hey chat, we haven't forgotten about you 👀
English
1.5K
167
8.9K
1.6M
N3crafter
N3crafter@n3crafter·
@TheRealAdamG We need validation, review, iteration meta layers. There is a lot of focus on generating code right now
English
0
0
1
74
Adam.GPT
Adam.GPT@TheRealAdamG·
Not sure if you heard... but we're building a superapp
Adam.GPT tweet media
English
31
24
423
21.4K