Bartosz Miselis @bmiselis

Solopreneur. DataOps engineer. I write about AI, data, cloud, and automation. toolongautomated.com planet Earth Joined October 2010

Tweets

816
Followers

121
Following

98
Likes

4K

Bartosz Miselis @bmiselis

6 months ago

@tomik99 Z ciekawości, jak zarobi na tym Open Mercato? Wdrożenia oparte na open source’owym sofcie?

0 0 0 234 0

View Details

Bartosz Miselis @bmiselis

7 months ago

it definitely feels different, surprisingly smart and agentic

Mckay Wrigley @mckaywrigley

7 months ago

@iruletheworldmo opus 4.5 might as well be agi for devs

21 12 318 175K 45

0 0 0 26 0

View Details

Bartosz Miselis @bmiselis

11 months ago

I mean, seems fine, even without Thinking explicitly enabled.

0 0 0 33 0

View Details

Bartosz Miselis @bmiselis

11 months ago

@emollick Seems like they could simply discard „minimal” effort 😅

0 0 0 126 0

View Details

@emollick It feels 50% of time is low probability, but how many times do software engineers spend X hours on something to discover they wasted the time entirely, of course without any solution? 50% is pretty high tbh

0 0 1 71 0

View Details

Bartosz Miselis @bmiselis

11 months ago

GPT-5 personalities are great. They allow to view certain events from differing perspective (e.g. „nerd” for facts gathering and „cynic” for more human-like judgement). I wish it was possible to ask a question to both at the same time and see them side by side for comparison.

0 0 1 60 0

View Details

Bartosz Miselis @bmiselis

11 months ago

@DeryaTR_ Superior to o3-pro? How would you measure it? Vibe check? Or factuality? Depth of insights?

1 0 2 391 0

View Details

Bartosz Miselis @bmiselis

11 months ago

oh boy

OpenAI @OpenAI

11 months ago

LIVE5TREAM THURSDAY 10AM PT

2K 3K 23K 6.9M 2K

0 0 1 43 0

View Details

Bartosz Miselis @bmiselis

11 months ago

@brandonwilson I wish I knew, Brandon

1 0 0 11 0

View Details

Bartosz Miselis @bmiselis

11 months ago

Really curious whether GPT-5 will simply be a subtle extension to the general UX of ChatGPT („drop the model picker” vibe), or whether it’ll introduce a new style of interaction.

Alexander Wei @alexwei_

11 months ago

8/N Btw, we are releasing GPT-5 soon, and we’re excited for you to try it. But just to be clear: the IMO gold LLM is an experimental research model. We don’t plan to release anything with this level of math capability for several months.

43 191 2K 523K 224

1 0 0 211 0

View Details

Bartosz Miselis @bmiselis

11 months ago

@mark_k @OpenAI depends on the project ;)

1 0 1 84 0

View Details

Bartosz Miselis @bmiselis

11 months ago

@daniel_mac8 @sama I think it's because they want to maintain enough competitive advantage.

0 0 0 243 0

View Details

Bartosz Miselis @bmiselis

12 months ago

@emollick I’m using it exactly the same way - describing things/products/spots I don’t know or remember the name of. It’s perfect in vast majority of places!

0 0 1 193 0

View Details

Bartosz Miselis @bmiselis

a year ago

@pierceboggan @code Quick file editing

0 0 0 78 0

View Details

Ethan Mollick @emollick

a year ago

New paper shows a familiar result on LLMs & medicine: Doctors given clinical vignettes produce significantly more accurate diagnoses when using a custom GPT built with the (obsolete) GPT-4 than doctors with Google/Pubmed but not AI. Yet AI alone is as accurate as doctors + AI.

30 186 967 403K 545

View Details

Bartosz Miselis @bmiselis

a year ago

@DeryaTR_ Forget o3-pro professor, GPT-5 hiding somewhere there!

1 0 7 444 2

View Details

Bartosz Miselis @bmiselis

a year ago

@arcprize @AnthropicAI I really like it that you’re benchmarking recent models, not only OpenAI’s!

0 0 0 326 0

View Details

Bartosz Miselis @bmiselis

a year ago

@omooretweets Or it’s the models gradually slowing down 😅

0 0 0 27 0

View Details

Bartosz Miselis @bmiselis

a year ago

@emollick Even so, it still feel like keeping a lion in a circus..

0 0 0 110 0

View Details

Bartosz Miselis @bmiselis

a year ago

I wish such conferences were shared with the public so that we can all hear about it as well :/

AI Notkilleveryoneism Memes ⏸️ @AISafetyMemes

a year ago

At an exclusive event of world leaders, Paul Tudor Jones says a top AI leader warned everyone: “It's going to take an accident where 50 to 100 million people die to make the world take the threat of this really seriously… I'm buying 100 acres in the Midwest, cattle, chickens."