I.Bogunov

PM, former software dev.

[I.Bogunov @tech_mind]

@memmaptensor Good luck, all through budget is kinda nuts for solo project.

[I.Bogunov @tech_mind]

@iikagenchan Давай еще ��онгридов,звучит интересно.

[laralara @FraLorik]

"За первый квартал падение 51%. Такого не было даже в 90-е»
Директор Череповецкого литейно-механического завода Боглаев заявил о катастрофическом обвале производства в сфере машиностроения.
При этом исправить ситуацию практически нереально, так как в бюджете нет денег для ⤵️

[⭐️Oles Filonenko⭐️ @FilonenkoOles]

Русский бизнес это конечно угар. 😂 Вы ебанулись штоле, что вы там делаете нахуй, бегите ахаха сссуккка 😂

[I.Bogunov @tech_mind]

@matthew_d_green All this shit is stupid as hell, what prohibits 10-14 year old from asking older brother/sister/friend to give him his phone or sign in to pc with his own account ? This is incredibly stupid.

[Luca Dellanna @DellAnnaLuca]

The reactions of many researchers on finally being held responsible for having read the very paper they submitted are... something.

[Luis @Thor_luisXVII]

Sinceramente ya no me importa salir bien en las fotos, con salir bien en las Resonancias y en las Radiografias me vale.

[GaryD20 Games @GaryD20Games]

Boobs win every time. 😋

[I.Bogunov @tech_mind]

@Shadow45201 What anime is this ?

[I.Bogunov @tech_mind]

@stalkermustang TLDR: code was specs and it's just higher level langauge now.

[I.Bogunov @tech_mind]

@stalkermustang Code was always the spec:just at a lower abstraction level.Machine code->assembler->C->Java/Python/frameworks->LLM instructions.The developer’s job was always to project vague product intent into millions of small decisions: DB, caching, data structures, edge cases, failure modes

[I.Bogunov @tech_mind]

@stalkermustang With agents, the question is who makes those decisions.If the LLM keeps asking the human, the human is still the developer, just prompting instead of coding.If the LLM makes good decisions itself, the human role shifts to intent, architecture, and review.

[Gary Marcus @GaryMarcus]

Some things never change. If you don’t understand this one, you don’t understand what’s happening AI.

Marcus, 1998: neural nets have trouble generalizing far beyond the data.

Marcus, 2001, 2012, 2019, 2022, etc: neural nets have trouble generalizing far beyond the data.

Apple, 2025: neural nets have trouble generalizing far beyond the data.

Meta/Stanford/Harvard, 2026: neural nets have trouble generalizing far beyond the data.

[Prorok @_Ilya_Prorok_]

путинский цирк шапито

[Deedy @deedydas]

The creators of SWE-Bench just dropped a really simple new benchmark every LLM gets 0% on.

ProgramBench asks: can models recreate real executable programs (ffmpeg, SQLite, ripgrep) from scratch with no internet?

We are far from saturated on model quality.

[I.Bogunov @tech_mind]

@RealPezduza Тот момент когда уже опасаешсяь лайкать, т.к. не хочешь чтобы расспростринлось.

[Zephyr @zephyr_z9]

So, Jensen was right all along...

[I.Bogunov @tech_mind]

@howindespair О, я документалку на ютубе про такое же смотрел в китае!