Ersel Gökmen

Co-Founder @ Buy Buddy

[Ersel Gökmen @ErselGokmen]

@ajambrosino Your business model first, then your product roadmap.

[Ersel Gökmen @ErselGokmen]

@levie Maybe we’re thinking about this too human-centrically. Today, humans build and deploy AI for companies. Tomorrow, AI systems may create demand themselvesfinding, hiring and paying humans for tasks they can’t do. What jobs will AI itself need humans to do?

[Ersel Gökmen @ErselGokmen]

Persistent cross-chat agents with local/cloud execution.

It would be great to have one Codex chat act as an orchestrator that can invoke agents living in other chats, while each agent keeps its own context and specialization.

More importantly, each agent should be able to run either locally or in the cloud. This would let us offload resource-heavy agents to cloud machines when local RAM or compute becomes a bottleneck, while keeping lightweight agents running locally.

For example: orchestrator and frontend agents locally, while backend, testing, or heavy build agents run instantly in the cloud all coordinated from the same chat.

[Hardik Gohil @GohilHardy]

genuine question

if Cursor already gives you

every Claude model

every Codex model

every Gemini model

plus its own Composer

why are people still paying for Claude or Codex separately?

what am I missing?

[AI at Meta @AIatMeta]

We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2.

Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication.

We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating.

🧵👇

[Ersel Gökmen @ErselGokmen]

@araseb_ End of simulation.

[Ersel Gökmen @ErselGokmen]

@gdb Six months ago, it felt like a helpful tool. Now it’s basically the coworker we talk to more than most humans.

[Ersel Gökmen @ErselGokmen]

https://t.co/Do182UDzNW

[Ersel Gökmen @ErselGokmen]

@garrytan Maybe the next layer after context selection is reading strategy.

Once the right docs are loaded, agents still need to skim like humans: scan the structure, find the relevant sections, then read deeply where it matters.

[Ersel Gökmen @ErselGokmen]

@satyanadella Agree. A frontier ecosystem may also create a new problem: agent fragmentation.

As every company and SaaS product builds its own agents, the opportunity may not be just agent communication, but coordination: shared context, tasks, memory, permissions, and accountability.

[Ersel Gökmen @ErselGokmen]

Agree. A frontier ecosystem may also create a new problem: agent fragmentation.

As every company and SaaS product builds its own agents, the opportunity may not be just agent communication, but coordination: shared context, tasks, memory, permissions, and accountability.

[Satya Nadella @satyanadella]

https://t.co/vLmiBKTtX3

[Ersel Gökmen @ErselGokmen]

@AliGrids It at least finally kill the illusion of "I thought about this idea before them."

[Ersel Gökmen @ErselGokmen]

@snowmaker Maybe people confuse “we don’t know how to make this work yet” with “this can’t be done.”

The interesting place to look might be ideas people clearly want, that seem possible in theory, but are still dismissed because nobody has made them work well enough yet.

[Ersel Gökmen @ErselGokmen]

Most models now, are trained for code.

A lot of business problems can be reduced into code problems.

But data science is not the same.

Code looks for a solution.

Data science has to look for what is true in the real world.

That requires grounding, context, and judgment.

[Ersel Gökmen @ErselGokmen]

Oh my god US government just one shotted Fable.

[Ersel Gökmen @ErselGokmen]

@aiseomastery @SantiTorAI Let me know I can give you some free credits 👍.

[Ersel Gökmen @ErselGokmen]

@rabi_guha We have solved that problem with "blocks" at https://t.co/0bymRKPP6t. Agent has pre created blocks and it just configures it. UI becomes consistent, lineage works 100%. You can connect any type of data.

[Ersel Gökmen @ErselGokmen]

@aiseomastery @SantiTorAI Try https://t.co/0bymRKPP6t
Not 30 seconds but does the job 👍

[Ersel Gökmen @ErselGokmen]

Apple’s new iOS 26 design isn’t made for phones—it’s made for mixed reality.
The glass-like UI that bends light and reacts to surroundings makes more sense through glass.
Like iOS 7 prepped us for flat design, this is Apple shaping what’s next.

[Ersel Gökmen @ErselGokmen]

@signulll Most AI agents will essentially be SaaS solutions wrapped with LLMs.